Academic Commons Search Results
https://academiccommons.columbia.edu/catalog?action=index&controller=catalog&f%5Bsubject_facet%5D%5B%5D=Operations+research&format=rss&fq%5B%5D=has_model_ssim%3A%22info%3Afedora%2Fldpd%3AContentAggregator%22&q=&rows=500&sort=record_creation_date+desc
Academic Commons Search Resultsen-usDynamic Trading Strategies in the Presence of Market Frictions
https://academiccommons.columbia.edu/catalog/ac:dv41ns1rpx
Saglam, Mehmet10.7916/D8KS7ZCDTue, 05 Dec 2017 17:29:20 +0000This thesis studies the impact of various fundamental frictions in the microstructure of financial markets. Specific market frictions we consider are latency in high-frequency trading, transaction costs arising from price impact or commissions, unhedgeable inventory risks due to stochastic volatility and time-varying liquidity costs. We explore the implications of each of these frictions in rigorous theoretical models from an investor's point of view and derive analytical expressions or efficient computational procedures for dynamic strategies. Specific methodologies in computing these policies include stochastic control theory, dynamic programming and tools from applied probability and stochastic processes.
In the first chapter, we describe a theoretical model for the quantitative valuation of latency and its impact on the optimal dynamic trading strategy. Our model measures the trading frictions created by the presence of latency, by considering the optimal execution problem of a representative investor. Via a dynamic programming analysis, our model provides a closed-form expression for the cost of latency in terms of well-known parameters of the underlying asset. We implement our model by estimating the latency cost incurred by trading on a human time scale. Examining NYSE common stocks from 1995 to 2005 shows that median latency cost across our sample more than tripled during this time period.
In the second chapter, we provide a highly tractable dynamic trading policy for portfolio choice problems with return predictability and transaction costs. Our rebalancing rule is a linear function of the return predicting factors and can be utilized in a wide spectrum of portfolio choice models with minimal assumptions. Linear rebalancing rules enable to compute exact and efficient formulations of portfolio choice models with linear constraints, proportional and nonlinear transaction costs, and quadratic utility function on the terminal wealth. We illustrate the implementation of the best linear rebalancing rule in the context of portfolio execution with positivity constraints in the presence of short-term predictability. We show that there exists a considerable performance gain in using linear rebalancing rules compared to static policies with shrinking horizon or a dynamic policy implied by the solution of the dynamic program without the constraints.
Finally, in the last chapter, we propose a factor-based model that incorporates common factor shocks for the security returns. Under these realistic factor dynamics, we solve for the dynamic trading policy in the class of linear policies analytically. Our model can accommodate stochastic volatility and liquidity costs as a function of factor exposures. Calibrating our model with empirical data, we show that our trading policy achieves superior performance in the presence of common factor shocks.Finance, Operations research, Stochastic processes, BusinessBusinessThesesEfficient Simulation Methods for Estimating Risk Measures
https://academiccommons.columbia.edu/catalog/ac:w3r2280gdd
Du, Yiping10.7916/D8J10FQ4Thu, 26 Oct 2017 13:25:16 +0000In this thesis, we analyze the computational problem of estimating financial risk in nested Monte Carlo simulation. An outer simulation is used to generate financial scenarios, and an inner simulation is used to estimate future portfolio values in each scenario. Mean squared error (MSE) for standard nested simulation converges at the rate $k^{-2/3}$, where $k$ is the computational budget.
In the first part of this thesis, we focus on one risk measure, the probability of a large loss, and we propose a new algorithm to estimate this risk. Our algorithm sequentially allocates computational effort in the inner simulation based on marginal changes in the risk estimator in each scenario. Theoretical results are given to show that the risk estimator has an asymptotic MSE of order $k^{-4/5+\epsilon}$, for all positive $\epsilon$, that is faster compared to the conventional uniform inner sampling approach. Numerical results consistent with the theory are presented.
In the second part of this thesis, we introduce a regression-based nested Monte Carlo simulation method for risk estimation. The proposed regression method combines information from different risk factor realizations to provide a better estimate of the portfolio loss function. The MSE of the regression method converges at the rate $k^{-1}$ until reaching an asymptotic bias level which depends on the magnitude of the regression error. Numerical results consistent with our theoretical analysis are provided and numerical comparisons with other methods are also given.
In the third part of this thesis, we propose a method based on weighted regression. Similar to the unweighted regression method, the MSE of the weighted regression method converges at the rate $k^{-1}$ until reaching an asymptotic bias level, which depends on the size of the regression error. However, the weighted approach further reduces MSE by emphasizing scenarios that are more important to the calculation of the risk measure. We find a globally optimal weighting strategy for general risk measures in an idealized setting. For applications, we propose and test a practically implementable two-pass method, where the first pass uses an unweighted regression and the second pass uses weights based on the first pass.Operations research, Monte Carlo method, Simulation methods, Risk assessmentIndustrial Engineering and Operations ResearchThesesInteger programming techniques for Polynomial Optimization
https://academiccommons.columbia.edu/catalog/ac:p8cz8w9gm0
Munoz, Gonzalo10.7916/D82F812GFri, 13 Oct 2017 22:18:27 +0000Modern problems arising in many domains are driving a need for more capable, state-of-the-art optimization tools. A sharp focus on performance and accuracy has appeared, for example, in science and engineering applications. In particular, we have seen a growth in studies related to Polynomial Optimization: a field with beautiful and deep theory, offering flexibility for modeling and high impact in diverse areas.
The understanding of structural aspects of the feasible sets in Polynomial Optimization, mainly studied in Real Algebraic Geometry, has a long tradition in Mathematics and it has recently acquired increased computational maturity, opening the gate for new Optimization methodologies to be developed. The celebrated hierarchies due to Lasserre, for example, emerged as good algorithmic templates. They allow the representation of semi-algebraic sets, possibly non-convex, through convex sets in lifted spaces, thus enabling the use of long-studied Convex Optimization methods. Nonetheless, there are some computational drawbacks for these approaches: they often rely on possibly large semidefinite programs, and due to scalability and numerical issues associated with SDPs, alternatives and complements are arising.
In this dissertation, we will explore theoretical and practical Integer-Programming-based techniques for Polynomial Optimization problems. We first present a Linear Programming relaxation for the AC-OPF problem in Power Systems, a non-convex quadratic problem, and show how such relaxation can be used to develop a tractable MIP-based algorithm for the AC Transmission Switching problem. From a more theoretical perspective, and motivated by the AC-OPF problem, we study how sparsity can be exploited as a tool for analysis of the fundamental complexity of a Polynomial Optimization problem, by showing LP formulations that can efficiently approximate sparse polynomial problems. Finally, we show a computationally practical approach for constructing strong LP approximations on-the-fly, using cutting plane approaches. We will show two different frameworks that can generate cutting planes, which are based on classical methods used in Mixed-Integer Programming.
Our methods mainly rely on the maturity of current MIP technology; we believe these contributions are important for the development of manageable approaches to general Polynomial Optimization problems.Operations research, Mathematics, Polynomials, Integer programming, Mathematical optimizationgm2543Industrial Engineering and Operations ResearchThesesAdaptive Algorithms for Ordinal Optimisation and Dynamic Pricing in E-commerce
https://academiccommons.columbia.edu/catalog/ac:66t1g1jwtr
Shin, Dongwook10.7916/D8154VM0Wed, 11 Oct 2017 22:18:39 +0000In Chapters 2 and 3, given a finite number of populations, henceforth referred to as systems, we are concerned with the problem of dynamically learning the statistical characteristics of the systems to ultimately select the best system. This is an instance of ordinal optimization where probability distributions governing each system's performance are not known, but can be learned via sequential sampling.
In Chapter 2 we study the classical setting where the ultimate goal is to choose the system with the highest mean, while in Chatper 3 the systems are compared based on quantiles. The latter setting is appropriate when downside or upside risk is more crucial than the mean performance. In both settings, we use large deviations theory to derive key structural insights on near-optimal allocation of the sampling budget, which are leveraged to design dynamic sampling policies that are practically implementable. We rigorously provide (asymptotic) performance guarantees for these policies. Further, we show via numerical testing that the proposed (nonparametric) policies perform competitively compared to other benchmark policies.
In Chapter 4, we investigate how the presence of product reviews affects a dynamic-pricing monopolist. A salient feature of our problem is that the demand function evolves over time in conjunction with the dynamics of the review system. The monopolist strives to maximize its total expected revenue over a finite horizon by adjusting prices in response to the review dynamics. To formulate the problem in tractable form, we study a fluid model, which is a good approximation when the volume of sales is large. This formulation lends itself to key structural insights, which are leveraged to design a well-performing pricing policy for the underlying revenue maximization problem. The proposed policy allows a closed-form expression for price and its performance is asymptotically near-optimal. We show via simulation and counterfactual analysis the effectiveness of the proposed policy in online markets with product reviews.Operations research, Mathematical optimization, Pricing, Electronic commerce, Algorithmsds2968BusinessThesesNon-Bayesian Inference and Prediction
https://academiccommons.columbia.edu/catalog/ac:v9s4mw6mcg
Xiao, Di10.7916/D8Q81RMVTue, 10 Oct 2017 22:16:52 +0000In this thesis, we first propose a coherent inference model that is obtained by distorting the prior density in Bayes' rule and replacing the likelihood with a so-called pseudo-likelihood. This model includes the existing non-Bayesian inference models as special cases and implies new models of base-rate neglect and conservatism. We prove a sufficient and necessary condition under which the coherent inference model is processing consistent, i.e., implies the same posterior density however the samples are grouped and processed retrospectively. We show that processing consistency does not imply Bayes' rule by proving a sufficient and necessary condition under which the coherent inference model can be obtained by applying Bayes' rule to a false stochastic model. We then propose a prediction model that combines a stochastic model with certain parameters and a processing-consistent, coherent inference model. We show that this prediction model is processing consistent, which states that the prediction of samples does not depend on how they are grouped and processed prospectively, if and only if this model is Bayesian. Finally, we apply the new model of conservatism to a car selection problem, a consumption-based asset pricing model, and a regime-switching asset pricing model.Operations research, Mathematical statistics, Industrial engineeringdx2125Industrial Engineering and Operations ResearchThesesOnline Algorithms for Dynamic Resource Allocation Problems
https://academiccommons.columbia.edu/catalog/ac:cnp5hqbzmv
Wang, Xinshang10.7916/D85H7TR6Fri, 29 Sep 2017 22:19:36 +0000Dynamic resource allocation problems are everywhere. Airlines reserve flight seats for those who purchase flight tickets. Healthcare facilities reserve appointment slots for patients who request them. Freight carriers such as motor carriers, railroad companies, and shipping companies pack containers with loads from specific origins to destinations.
We focus on optimizing such allocation problems where resources need to be assigned to customers in real time. These problems are particularly difficult to solve because they depend on random external information that unfolds gradually over time, and the number of potential solutions is overwhelming to search through by conventional methods.
In this dissertation, we propose viable allocation algorithms for industrial use, by fully leveraging data and technology to produce gains in efficiency, productivity, and usability of new systems. The first chapter presents a summary of major methodologies used in modeling and algorithm design, and how the methodologies are driven by the size of accessible data.
Chapters 2 to 5 present genuine research results of resource allocation problems that are based on Wang and Truong (2017); Wang et al. (2015); Stein et al. (2017); Wang et al. (2016). The algorithms and models cover problems in multiple industries, from a small clinic that aims to better utilize its expensive medical devices, to a technology giant that needs a cost-effective, distributed resource-allocation algorithm in order to maintain the relevance of its advertisements to hundreds of millions of consumers.Operations research, Resource allocation, Algorithmsxw2230Industrial Engineering and Operations ResearchThesesClearinghouse Default Resources: Theory and Empirical Analysis
https://academiccommons.columbia.edu/catalog/ac:hmgqnk98v6
Cheng, Wan-Schwin Allen10.7916/D8CN7GCCTue, 26 Sep 2017 22:27:53 +0000Clearinghouses insure trades. Acting as a central counterparty (CCP), clearinghouses consolidate financial exposures across multiple institutions, aiding the efficient management of counterparty credit risk. In this thesis, we study the decision problem faced by for-profit clearinghouses, focusing on primary economic incentives driving their determination of layers of loss-absorbing capital. The clearinghouse's loss-allocation mechanism, referred to as the default waterfall, governs the allocation and management of counterparty risk. This stock of loss-absorbing capital typically consists of initial margins, default funds, and the clearinghouse's contributed equity.
We separate the overall decision problem into two distinct subproblems and study them individually. The first is the clearinghouse's choice of initial margin and clearing fee requirements, and the second involves its choice of resources further down the waterfall, namely the default funds and clearinghouse equity. We solve for the clearinghouse's equilibrium choices in both cases explicitly, and address the different economic roles they play in the clearinghouse's profit-maximization process.
The models presented in this thesis show, without exception, that clearinghouse choices should depend not only on the riskiness of the cleared position but also on market and participants' characteristics such as default probabilities, fundamental value, and funding opportunity cost.
Our results have important policy implications. For instance, we predict a counteracting force that dampens monetary easing enacted via low interest rate policies. When funding opportunity costs are low, our research shows that clearinghouses employ highly conservative margin and default funds, which tie up capital and credit. This is supported by the low interest rate environment following the financial crisis of 2007--08. In addition to low productivity growth and return on capital, major banks have chosen to accumulate large cash piles on their balance sheets rather than increase lending. In terms of systemic risk, our empirical work, joint with the U.S. Commodity Futures Trading Commission (CFTC), points to the possibility of destabilizing loss and margin spirals: in the terminology of Brunnermeier and Pedersen (2009), we argue that a major clearinghouse's behavior is consistent with that of an uninformed financier and that common shocks to credit quality can lead to tightening margin constraints.Finance, Clearinghouses (Banking), Statistical decision, Operations researchwc2232Industrial Engineering and Operations ResearchThesesFundamental Tradeoffs for Modeling Customer Preferences in Revenue Management
https://academiccommons.columbia.edu/catalog/ac:dncjsxkspb
Desir, Antoine Minh10.7916/D8F76QZ0Wed, 02 Aug 2017 16:11:46 +0000Revenue management (RM) is the science of selling the right product, to the right person, at the right price. A key to the success of RM, which now spans a broad array of industries, is its grounding in mathematical modeling and analytics. This dissertation contributes to the development of new RM tools by: (1) exploring some fundamental tradeoffs underlying any RM problems, and (2) designing efficient algorithms for some RM applications. Another underlying theme of this dissertation is the modeling of customer preferences, a key component of any RM problem.
The first chapters of this dissertation focus on the model selection problem: many demand models are available but picking the right model is a challenging task. In particular, we explore the tension between the richness of a model and its tractability. To quantify this tradeoff, we focus on the assortment optimization problem, a very general and core RM problem. To capture customer preferences in this context, we use choice models, a particular type of demand model. In Chapters 1, 2, 3 and 4 we design efficient algorithms for the assortment optimization problem under different choice models. By assessing the strengths and weaknesses of different choice models, we can quantify the cost in tractability one has to pay for better predictive power. This in turn leads to a better understanding of the tradeoffs underlying the model selection problem.
In Chapter 5, we focus on a different question underlying any RM problem: choos- ing how to sell a given product. We illustrate this tradeoff by focusing on the problem of selling ad impressions via Internet display advertising platforms. In particular, we study how the presence of risk-averse buyers affects the desire for reservation con- tracts over real time buy via a second-price auction. In order to capture the risk aversion of buyers, we study different utility models.Operations research, Revenue management, Consumers' preferences--Mathematical models, Industrial engineeringad2918Industrial Engineering and Operations ResearchThesesOptimal Dynamic Strategies for Index Tracking and Algorithmic Trading
https://academiccommons.columbia.edu/catalog/ac:xwdbrv15h9
Ward, Brian Michael10.7916/D82F80R1Fri, 21 Jul 2017 19:11:27 +0000In this thesis we study dynamic strategies for index tracking and algorithmic trading. Tracking problems have become ever more important in Financial Engineering as investors seek to precisely control their portfolio risks and exposures over different time horizons. This thesis analyzes various tracking problems and elucidates the tracking errors and strategies one can employ to minimize those errors and maximize profit.
In Chapters 2 and 3, we study the empirical tracking properties of exchange traded funds (ETFs), leveraged ETFs (LETFs), and futures products related to spot gold and the Chicago Board Option Exchange (CBOE) Volatility Index (VIX), respectively. These two markets provide interesting and differing examples for understanding index tracking. We find that static strategies work well in the nonleveraged case for gold, but fail to track well in the corresponding leveraged case. For VIX, tracking via neither ETFs, nor futures portfolios succeeds, even in the nonleveraged case. This motivates the need for dynamic strategies, some of which we construct in these two chapters and further expand on in Chapter 4. There, we analyze a framework for index tracking and risk exposure control through financial derivatives. We derive a tracking condition that restricts our exposure choices and also define a slippage process that characterizes the deviations from the index over longer horizons. The framework is applied to a number of models, for example, Black-Scholes model and Heston model for equity index tracking, as well as the Square Root (SQR) model and the Concatenated Square Root (CSQR) model for VIX tracking. By specifying how each of these models fall into our framework, we are able to understand the tracking errors in each of these models.
Finally, Chapter 5 analyzes a tracking problem of a different kind that arises in algorithmic trading: schedule following for optimal execution. We formulate and solve a stochastic control problem to obtain the optimal trading rates using both market and limit orders. There is a quadratic terminal penalty to ensure complete liquidation as well as a trade speed limiter and trader director to provide better control on the trading rates. The latter two penalties allow the trader to tailor the magnitude and sign (respectively) of the optimal trading rates. We demonstrate the applicability of the model to following a benchmark schedule. In addition, we identify conditions on the model parameters to ensure optimality of the controls and finiteness of the associated value functions. Throughout the chapter, numerical simulations are provided to demonstrate the properties of the optimal trading rates.Operations research, Mathematics, Finance, Financial engineeringbmw2150Industrial Engineering and Operations ResearchThesesStructured Tensor Recovery and Decomposition
https://academiccommons.columbia.edu/catalog/ac:x3ffbg79fs
Mu, Cun10.7916/D8DV1X6MMon, 17 Jul 2017 16:14:19 +0000Tensors, a.k.a. multi-dimensional arrays, arise naturally when modeling higher-order objects and relations. Among ubiquitous applications including image processing, collaborative filtering, demand forecasting and higher-order statistics, there are two recurring themes in general: tensor recovery and tensor decomposition. The first one aims to recover the underlying tensor from incomplete information; the second one is to study a variety of tensor decompositions to represent the array more concisely and moreover to capture the salient characteristics of the underlying data. Both topics are respectively addressed in this thesis.
Chapter 2 and Chapter 3 focus on low-rank tensor recovery (LRTR) from both theoretical and algorithmic perspectives. In Chapter 2, we first provide a negative result to the sum of nuclear norms (SNN) model---an existing convex model widely used for LRTR; then we propose a novel convex model and prove this new model is better than the SNN model in terms of the number of measurements required to recover the underlying low-rank tensor. In Chapter 3, we first build up the connection between robust low-rank tensor recovery and the compressive principle component pursuit (CPCP), a convex model for robust low-rank matrix recovery. Then we focus on developing convergent and scalable optimization methods to solve the CPCP problem. In specific, our convergent method, proposed by combining classical ideas from Frank-Wolfe and proximal methods, achieves scalability with linear per-iteration cost.
Chapter 4 generalizes the successive rank-one approximation (SROA) scheme for matrix eigen-decomposition to a special class of tensors called symmetric and orthogonally decomposable (SOD) tensor. We prove that the SROA scheme can robustly recover the symmetric canonical decomposition of the underlying SOD tensor even in the presence of noise. Perturbation bounds, which can be regarded as a higher-order generalization of the Davis-Kahan theorem, are provided in terms of the noise magnitude.Operations research, Computer science, Statistics, Calculus of tensorscm3052Industrial Engineering and Operations ResearchThesesProduction Planning with Risk Hedging
https://academiccommons.columbia.edu/catalog/ac:tdz08kprt5
Wang, Liao10.7916/D8MP5FMVTue, 11 Jul 2017 01:07:55 +0000We study production planning integrated with risk hedging in a continuous-time stochastic setting. The (cumulative) demand process is modeled as a sum of two components: the demand rate is a general function in a tradable financial asset (which follows another stochastic process), and the noise component follows an independent Brownian motion. There are two decisions: a production quantity decision at the beginning of the planning horizon, and a dynamic hedging strategy throughout the horizon. Thus, the total terminal wealth has two components: production payoff, and profit/loss from the hedging strategy.
The production quantity and hedging strategy are jointly optimized under the mean-variance and the shortfall criteria. For each risk objective, we derive the optimal hedging strategy in closed form and express the associated minimum risk as a function of the production quantity, the latter is then further optimized. With both production and hedging (jointly) optimized, we provide a complete characterization of the efficient frontier. By quantifying the risk reduction contributed by the hedging strategy, we demonstrate its substantial improvement over a production-only decision.
To derive the mean-variance hedging strategy, we use a numeraire-based approach, and the derived optimal strategy consists of a risk mitigation component and an investment component. For the shortfall hedging, a convex duality method is used, and the optimal strategy takes the form of a put option and a digital option, which combine to close the gap from the target left by production (only).
Furthermore, we extend the models and results by allowing multiple products, with demand rates depending on multiple assets. We also make extension by allowing the asset price to follow various stochastic processes (other than the geometric Brownian motion).Operations researchlw2489Industrial Engineering and Operations ResearchThesesSoft Regulation with Crowd Recommendation: Coordinating Self-Interested Agents in Sociotechnical Systems under Imperfect Information
https://academiccommons.columbia.edu/catalog/ac:197950
Luo, Yu; Iyengar, Garud N.; Venkatasubramanian, Venkat10.7916/D8HM58FXFri, 30 Jun 2017 16:57:24 +0000Regulating emerging industries is challenging, even controversial at times. Under-regulation can result in safety threats to plant personnel, surrounding communities, and the environment. Over-regulation may hinder innovation, progress, and economic growth. Since one typically has limited understanding of, and experience with, the novel technology in practice, it is difficult to accomplish a properly balanced regulation. In this work, we propose a control and coordination policy called soft regulation that attempts to strike the right balance and create a collective learning environment. In soft regulation mechanism, individual agents can accept, reject, or partially accept the regulator’s recommendation. This non-intrusive coordination does not interrupt normal operations. The extent to which an agent accepts the recommendation is mediated by a confidence level (from 0 to 100%). Among all possible recommendation methods, we investigate two in particular: the best recommendation wherein the regulator is completely informed and the crowd recommendation wherein the regulator collects the crowd’s average and recommends that value. We show by analysis and simulations that soft regulation with crowd recommendation performs well. It converges to optimum, and is as good as the best recommendation for a wide range of confidence levels. This work sheds a new theoretical perspective on the concept of the wisdom of crowds.Learning, Operations research, Collective behavior, Sociotechnical systems, Sociology, Industrial engineering, System theoryyl2750, gi10, vv2213Chemical Engineering, Industrial Engineering and Operations ResearchArticlesEvacuating Damaged and Destroyed Buildings on 9/11: Behavioral and Structural Barriers
https://academiccommons.columbia.edu/catalog/ac:174721
Stellman, Steven D.; Kravitt, Alexandra; Groeger, Justina L.; Brackbill, Robert M.10.7916/D8DB7ZXJWed, 28 Jun 2017 20:28:26 +0000Introduction
Evacuation of the World Trade Center (WTC) twin towers and surrounding buildings damaged in the September 11, 2001 attacks provides a unique opportunity to study factors that affect emergency evacuation of high rise buildings.
Problem
The goal of this study is to understand the extent to which structural and behavioral barriers and limitations of personal mobility affected evacuation by occupants of affected buildings on September 11, 2001.
Methods
This analysis included 5,023 civilian, adult enrollees within the World Trade Center Health Registry who evacuated the two World Trade Center towers and over 30 other Lower Manhattan buildings that were damaged or destroyed on September 11, 2001. Multinomial logistic regression was used to predict total evacuation time (less than 30 to ≤60 minutes, greater than 1 hour to less than 2 hours relative to ≤30 minutes) in relation to number of infrastructure barriers and number of behavioral barriers, adjusted for demographic and other factors.
Results
A higher percentage of evacuees reported encountering at least one behavioral barrier (84.9%) than reported at least one infrastructure barrier (51.9%). This pattern was consistent in all buildings except WTC 1, the first building attacked, where greater than 90% of evacuees reported encountering both types of barriers. Smoke and poor lighting were the most frequently-reported structural barriers. Extreme crowding, lack of communication with officials, and being surrounded by panicked crowds were the most frequently-reported behavioral barriers. Multivariate analyses showed evacuation time to be independently associated with the number of each type of barrier as well as gender (longer times for women), but not with the floor from which evacuation began. After adjustment, personal mobility impairment was not associated with increased evacuation time.
Conclusion
Because most high-rise buildings have unique designs, infrastructure factors tend to be less predictable than behavioral factors, but both need to be considered in developing emergency evacuation plans in order to decrease evacuation time and, consequently, risk of injury and death during an emergency evacuation.Psychology, Operations researchsds91EpidemiologyArticlesChance Constrained Optimal Power Flow: Risk-Aware Network Control under Uncertainty
https://academiccommons.columbia.edu/catalog/ac:156182
Bienstock, Daniel; Chertkov, Michael; Harnett, Sean10.7916/D8VH5ZQVTue, 27 Jun 2017 17:57:54 +0000When uncontrollable resources fluctuate, Optimum Power Flow (OPF), routinely used by the electric power industry to re-dispatch hourly controllable generation (coal, gas and hydro plants) over control areas of transmission networks, can result in grid instability, and, potentially, cascading outages. This risk arises because OPF dispatch is computed without awareness of major uncertainty, in particular fluctuations in renewable output. As a result, grid operation under OPF with renewable variability can lead to frequent conditions where power line flow ratings are significantly exceeded. Such a condition, which is borne by simulations of real grids, would likely resulting in automatic line tripping to protect lines from thermal stress, a risky and undesirable outcome which compromises stability. Smart grid goals include a commitment to large penetration of highly fluctuating renewables, thus calling to reconsider current practices, in particular the use of standard OPF. Our Chance Constrained (CC) OPF corrects the problem and mitigates dangerous renewable fluctuations with minimal changes in the current operational procedure. Assuming availability of a reliable wind forecast parameterizing the distribution function of the uncertain generation, our CC-OPF satisfies all the constraints with high probability while simultaneously minimizing the cost of economic re-dispatch. CC-OPF allows efficient implementation, e.g. solving a typical instance over the 2746-bus Polish network in 20 seconds on a standard laptop.Industrial engineering, Operations researchdb17, srh2144Industrial Engineering and Operations ResearchArticlesChance Constrained Optimal Power Flow: Risk-Aware Network Control under Uncertainty
https://academiccommons.columbia.edu/catalog/ac:153902
Bienstock, Daniel; Chertkov, Michael; Harnett, Sean10.7916/D8PN9GCXTue, 27 Jun 2017 15:37:47 +0000When uncontrollable resources fluctuate, Optimum Power Flow (OPF), routinely used by the electric power industry to redispatch hourly controllable generation (coal, gas and hydro plants) over control areas of transmission networks, can result in grid instability, and, potentially, cascading outages. This risk arises because OPF dispatch is computed without awareness of major uncertainty, in particular fluctuations in renewable output. As a result, grid operation under OPF with renewable variability can lead to frequent conditions where power line flow ratings are significantly exceeded. Such a condition, which is borne by simulations of real grids, would likely resulting in automatic line tripping to protect lines from thermal stress, a risky and undesirable outcome which compromises stability. Smart grid goals include a commitment to large penetration of highly fluctuating renewables, thus calling to reconsider current practices, in particular the use of standard OPF. Our Chance Constrained (CC) OPF corrects the problem and mitigates dangerous renewable fluctuations with minimal changes in the current operational procedure. Assuming availability of a reliable wind forecast parameterizing the distribution function of the uncertain generation, our CCOPF satisfies all the constraints with high probability while simultaneously minimizing the cost of economic redispatch. CCOPF allows efficient implementation, e.g. solving a typical instance over the 2746bus Polish network in 20s on a standard laptop.Industrial engineering, Operations researchdb17, srh2144Industrial Engineering and Operations ResearchArticlesOptimization in Strategic Environments
https://academiccommons.columbia.edu/catalog/ac:202863
Feigenbaum, Itai Izhak10.7916/D8377916Thu, 15 Jun 2017 16:10:42 +0000This work considers the problem faced by a decision maker (planner) trying to optimize over incomplete data. The missing data is privately held by agents whose objectives are dierent from the planner's, and who can falsely report it in order to advance their objectives. The goal is to design optimization mechanisms (algorithms) that achieve "good" results when agents' reports follow a game-theoretic equilibrium. In the first part of this work, the goal is to design mechanisms that provide a small worst-case approximation ratio (guarantee a large fraction of the optimal value in all instances) at equilibrium. The emphasis is on strategyproof mechanisms|where truthfulness is a dominant strategy equilibrium|and on the approximation ratio at that equilibrium. Two problems are considered|variants of knapsack and facility location problems. In the knapsack problem, items are privately owned by agents, who can hide items or report fake ones; each agent's utility equals the total value of their own items included in the knapsack, while the planner wishes to choose the items that maximize the sum of utilities. In the facility location problem, agents have private linear single sinked/peaked preferences regarding the location of a facility on an interval, while the planner wishes to locate the facility in a way that maximizes one of several objectives. A variety of mechanisms and lower bounds are provided for these problems. The second part of this work explores the problem of reassigning students to schools. Students have privately known preferences over the schools. After an initial assignment is made, the students' preferences change, get reported again, and a reassignment must be obtained. The goal is to design a reassignment mechanism that incentivizes truthfulness, provides high student welfare, transfers relatively few students from their initial assignment, and respects student priorities at schools. The class of mechanisms considered is permuted lottery deferred acceptance (PLDA) mechanisms, which is a natural class of mechanisms based on permuting the lottery numbers students initially draw to decide the initial assignment. Both theoretical and experimental evidence is provided to support the use of a PLDA mechanism called reversed lottery deferred acceptance (RLDA). The evidence suggests that under some conditions, all PLDA mechanisms generate roughly equal welfare, and that RLDA minimizes transfers among PLDA mechanisms.Mathematical optimization, Algorithms, Planners, Operations researchiif2103Industrial Engineering and Operations ResearchThesesMulti-Agent Control in Sociotechnical Systems
https://academiccommons.columbia.edu/catalog/ac:208851
Luo, Yu10.7916/D8FX7G35Thu, 15 Jun 2017 16:08:15 +0000Process control is essential in chemical engineering and has diverse applications in automation, manufacturing, scheduling, etc. In this cross-disciplinary work, we shift the domain focus from the control of machines to the control of multiple intelligent agents. Our goal is to improve the optimization problem-solving process, such as optimal regulation of emerging technologies, in a multi-agent system. Achieving that improvement would have potential value both within and outside the chemical engineering community. This work also illustrates the possibility of applying process systems engineering techniques, especially process control, beyond chemical plants.
It is very common to observe crowds of individuals solving similar problems with similar information in a largely independent manner. We argue here that the crowds can become more efficient and robust problem-solvers, by partially following the average opinion. This observation runs counter to the widely accepted claim that the wisdom of crowds deteriorates with social influence. The key difference is that individuals are self-interested and hence will reject feedbacks that do not improve their performance. We propose a multi-agent control-theoretic methodology, soft regulation, to model the collective dynamics and compute the degree of social influence, i.e., the level to which one accepts the population feedback, that optimizes the problem-solving performance.
Soft regulation is a modeling language for multi-agent sociotechnical systems. The state-space formulation captures the individual learning process (i.e., open loop dynamics) as well as the influence of the population feedback in a straightforward manner. It can model a diverse set of existing multi-agent dynamics. Through numerical analysis and linear algebra, we attempt to understand the role of feedback in multi-agent collective dynamics, thus achieving multi-agent control in sociotechnical systems.
Our analysis through mathematical proofs, simulations, and a human subject experiment suggests that intelligent individuals, solving the same problem (or similar problems), could do much better by adaptively adjusting their decisions towards the population average. We even discover that the crowd of human subjects could self-organize into a near-optimal setting. This discovery suggests a new coordination mechanism for enhancing individual decision-making. Potential applications include mobile health, urban planning, and policymaking.Intellect, Sociotechnical systems, Multiagent systems, Chemical process control, Chemical engineering, Operations research, Mathematicsyl2750Chemical EngineeringThesesCompetition in Service Operations and Supply Chains: Equilibrium Analysis and Structural Estimation
https://academiccommons.columbia.edu/catalog/ac:202431
Lu, Lijian10.7916/D83N23MKThu, 15 Jun 2017 16:08:15 +0000The service industry has become increasingly competitive. This dissertation addresses a number of outstanding and fundamental questions of competitions in service operations and supply chains. The challenges are characterization of the equilibrium behaviors, estimating the impact of firms' interactions, and designing of efficient market mechanisms.
The first chapter of this dissertation considers price competition models for oligopolistic markets, in which the consumer reacts to relative rather than absolute prices, where the relative price is defined as the difference between the absolute price and a given reference value. Such settings arise, for example, when the full retail price earned by the ``retailer" is reduced by virtue of a third party offering a subsidy or a rebate or in prospect theoretical models in which customers establish a reference price and base their choices on the differentials with respect to the reference price. When choosing among the various competing options, the consumer trades off the net price paid with various other product or service attributes, as in standard price competition models. The reference price may be exogenously specified and pre-announced to the competing firms. Alternatively, it may be endogenously determined, as a function of the set of absolute prices selected by the competing firms, for example the lowest or the second lowest price. We characterize the equilibrium behavior under a general reference value scheme of the above type; this in a base model, where we assume that the consumer choice model is of the general MultiNomialLogit (MNL) type. We also derive comparison results for the price equilibria that arise under alternative subsidy schemes. These comparisons have important implications for the design of subsidy schemes.
The second chapter applies the results of the first chapter to the Medicare insurance market, both in terms of its existing structure, as well as in terms of various proposals to redesign the program. Based on an oligopoly price competition model tailored towards this market, and actual county-by-county data for the year 2010, we estimate the impact such reforms would have on the plans' market shares, equilibrium premia, the government's cost, and the out-of-pocket expenses of the beneficiaries. We employ two different methodologies to derive the parameters in the county-by-county competition models: (i) a calibration model, and (ii) parameter distributions obtained from models estimated in Curto et al. (2015). The predicted impacts on the above performance measures are remarkably consistent across the two methodologies and reveal, for example, that the government cost would decrease by 8% if the traditional fee-for-service(FFS) plans are kept out of competitive bidding process and by 16.5%-21% if they are part of the process.
The third chapter studies a class of buy procurement mechanisms, framework agreements (FAs), that are commonly used by buying agencies around the world to satisfy demand that arises over a certain time horizon. We are one of the first in the literature that provides a formal understanding of FAs, with a particular focus on the cost uncertainty faced by bidders over the FA time horizon. We introduce a model that generalizes standard auction models to include this salient feature of FAs; we analyze this model theoretically and numerically. First, we show that FAs are subject to a sort of winner's curse that in equilibrium induces higher expected buying prices relative to running first-price auctions as needs arise. Then, our results provide concrete design recommendations that alleviate this issue and decrease buying prices in FAs, highlighting the importance of (i) monitoring the price charged at the open market by the FA winner and using it to bound the buying price; (ii) investing in implementing price indexes for the random part of suppliers' costs; and (iii) allowing suppliers the flexibility to reduce their prices to compete with the open market throughout the selling period. These prescriptions are already being used by the Chilean government procurement agency that buys US$2 billion worth of contracts every year using FAs.
The fourth chapter considers the preference of contractual forms in supply chains. The supply chain contracting literature has focused on incentive contracts designed to align supply chain members' individual interests. A key finding of this literature is that members' preferences for contractual forms are often at odds: the upstream supplier prefers more complex contracts that can coordinate the supply chain; however, the downstream retailer prefers the wholesale price--only contract because it leaves more surplus (than a coordinating contract) that the retailer can get. This chapter addresses the following question: under what circumstances do suppliers and retailers prefer the same contractual form? We study supply chain members' preference for contractual forms in three different competitive settings in which multiple supply chains compete to sell substitutable products to the same market. Our analysis suggests that both upstream and downstream sides of the supply chains may prefer the same ``quantity discount'' contract, thereby eliminating the conflicts of interest that otherwise typify contracting situations. More interesting still is that both sides may also prefer the wholesale price--only contract, which offers a theoretical explanation to why the simple inefficient contract is widely adopted in supply chain transactions.Business logistics, Prices--Econometric models, Service industries, Operations research, Economics, Health services administration, Medical carell2755BusinessThesesOptimal Stopping and Switching Problems with Financial Applications
https://academiccommons.columbia.edu/catalog/ac:204576
Wang, Zheng10.7916/D8VQ330DThu, 15 Jun 2017 16:07:59 +0000This dissertation studies a collection of problems on trading assets and derivatives over finite and infinite horizons. In the first part, we analyze an optimal switching problem with transaction costs that involves an infinite sequence of trades. The investor's value functions and optimal timing strategies are derived when prices are driven by an exponential Ornstein-Uhlenbeck (XOU) or Cox-Ingersoll-Ross (CIR) process. We compare the findings to the results from the associated optimal double stopping problems and identify the conditions under which the double stopping and switching problems admit the same optimal entry and/or exit timing strategies. Our results show that when prices are driven by a CIR process, optimal strategies for the switching problems are of the classic buy-low-sell-high type. On the other hand, under XOU price dynamics, the investor should refrain from entering the market if the current price is very close to zero. As a result, the continuation (waiting) region for entry is disconnected. In both models, we provide numerical examples to illustrate the dependence of timing strategies on model parameters. In the second part, we study the problem of trading futures with transaction costs when the underlying spot price is mean-reverting. Specifically, we model the spot dynamics by the OU, CIR or XOU model. The futures term structure is derived and its connection to futures price dynamics is examined. For each futures contract, we describe the evolution of the roll yield, and compute explicitly the expected roll yield. For the futures trading problem, we incorporate the investor's timing options to enter and exit the market, as well as a chooser option to long or short a futures upon entry. This leads us to formulate and solve the corresponding optimal double stopping problems to determine the optimal trading strategies. Numerical results are presented to illustrate the optimal entry and exit boundaries under different models. We find that the option to choose between a long or short position induces the investor to delay market entry, as compared to the case where the investor pre-commits to go either long or short. Finally, we analyze the optimal risk-averse timing to sell a risky asset. The investor's risk preference is described by the exponential, power or log utility. Two stochastic models are considered for the asset price -- the geometric Brownian motion (GBM) and XOU models to account for, respectively, the trending and mean-reverting price dynamics. In all cases, we derive the optimal thresholds and certainty equivalents to sell the asset, and compare them across models and utilities, with emphasis on their dependence on asset price, risk aversion, and quantity. We find that the timing option may render the investor's value function and certainty equivalent non-concave in price even though the utility function is concave in wealth. Numerical results are provided to illustrate the investor's optimal strategies and the premia associated with optimally timing to sell with different utilities under different price dynamics.Assets (Accounting), Mathematical optimization, Ornstein-Uhlenbeck process, Brownian motion processes, Finance--Mathematical models, Operations research, Financezw2192Industrial Engineering and Operations ResearchThesesApproximation Algorithms for Demand-Response Contract Execution and Coflow Scheduling
https://academiccommons.columbia.edu/catalog/ac:202248
Qiu, Zhen10.7916/D8FQ9WVPThu, 15 Jun 2017 15:06:18 +0000Solving operations research problems with approximation algorithms has been an important topic since approximation algorithm can provide near-optimal solutions to NP-hard problems while achieving computational efficiency. In this thesis, we consider two different problems in the field of optimal control and scheduling theory respectively and develop efficient approximation algorithms for those problems with performance guarantee.
Chapter 2 presents approximation algorithms for solving the optimal execution problem for demand-response contract in electricity markets. Demand side participation is essential for achieving real-time energy balance in today's electricity grid. Demand-response contracts, where an electric utility company buys options from consumers to reduce their load in the future, are an important tool to increase demand-side participation. In this chapter, we consider the operational problem of optimally exercising the available contracts over the planning horizon such that the total cost to satisfy the demand is minimized. In particular, we consider the objective of minimizing the sum of the expected ℓ_β-norm of the load deviations from given thresholds and the contract execution costs over the planning horizon. For β=∞, this reduces to minimizing the expected peak load. The peak load provides a good proxy to the total cost of the utility as spikes in electricity prices are observed only in peak load periods. We present a data driven near-optimal algorithm for the contract execution problem. Our algorithm is a sample average approximation (SAA) based dynamic program over a multi-period planning horizon. We provide a sample complexity bound on the number of demand samples required to compute a (1+ε)-approximate policy for any ε>0. Our SAA algorithm is quite general and we show that it can be adapted to quite general demand models including Markovian demands and objective functions. For the special case where the demand in each period is i.i.d., we show that a static solution is optimal for the dynamic problem. We also conduct a numerical study to compare the performance of our SAA based DP algorithm. Our numerical experiments show that we can achieve a (1+ε)-approximation in significantly smaller numbers of samples than what is implied by the theoretical bounds. Moreover, the structure of the approximate policy also shows that it can be well approximated by a simple affine function of the state.
In Chapter 3, we study the NP-hard coflow scheduling problem and develop a polynomial-time approximation algorithm for the problem with constant approximation ratio. Communications in datacenter jobs (such as the shuffle operations in MapReduce applications) often involve many parallel flows, which may be processed simultaneously. This highly parallel structure presents new scheduling challenges in optimizing job-level performance objectives in data centers. Chowdhury and Stoica [13] introduced the coflow abstraction to capture these communication patterns, and recently Chowdhury et al. [15] developed effective heuristics to schedule coflows. In this chapter, we consider the problem of efficiently scheduling coflows so as to minimize the total weighted completion time, which has been shown to be strongly NP-hard [15]. Our main result is the first polynomial-time deterministic approximation algorithm for this problem, with an approximation ratio of $64/3$, and a randomized version of the algorithm, with a ratio of 8+16sqrt{2}/3. Our results use techniques from both combinatorial scheduling and matching theory, and rely on a clever grouping of coflows.
In Chapter 4, we carry out a comprehensive experimental analysis on a Facebook trace and extensive simulated instances to evaluate the practical performance of several algorithms for coflow scheduling, including our approximation algorithms developed in Chapter 3. Our experiments suggest that simple algorithms provide effective approximations of the optimal, and that the performance of the approximation algorithm of Chapter 3 is relatively robust, near optimal, and always among the best compared with the other algorithms, in both the offline and online settings.Approximation algorithms, Mathematical optimization, Scheduling, Operations researchzq2110Industrial Engineering and Operations ResearchThesesTractable approximation algorithms for high dimensional sequential optimization problems,
https://academiccommons.columbia.edu/catalog/ac:195662
Bhat, Nikhil10.7916/D8JQ10W8Thu, 15 Jun 2017 15:04:39 +0000Sequential decision making problems are ubiquitous in a number of research areas such as operations research, finance, engineering and computer science. The main challenge with these problems comes from the fact that, firstly, there is uncertainty about the future. And secondly, decisions have to be made over a period of time, sequentially. These problems, in many cases, are modeled as Markov Decision Process (MDP). Most real-life MDPs are ‘high dimensional’ in nature making them challenging from a numerical point of view. We consider a number of such high dimensional MDPs. In some cases such problems can be approximately solved using Approximate Dynamic Programming. In other cases problem specific analysis can be solved to device tractable policies that are near-optimal. In Chapter 2, we presents a novel and practical non-parametric approximate dynamic programming (ADP) algorithm that enjoys graceful, dimension-independent approximation and sample complexity guarantees. In particular, we establish both theoretically and computationally that our proposal can serve as a viable replacement to state of the art parametric ADP algorithms, freeing the designer from carefully specifying an approximation architecture. We accomplish this by ‘kernelizing’ a recent mathematical program for ADP (the ‘smoothed’ approximate LP) proposed by [Desai et al., 2011]. In Chapter 3, we consider a class of stochastic control problems where the action space at each time can be described by a class of matching or, more generally, network flow polytopes. Special cases of this class of dynamic matching problems include many problems that are well-studied in the literature, such as: (i) online keyword matching in Internet advertising (the adwords problem); (ii) the bipartite matching of donated kidneys from cadavers to recipients; and (iii) the allocation of donated kidneys through exchanges over cycles of live donor-patient pairs. We provide an approximate dynamic program (ADP) algorithm for dynamic matching with stochastic arrivals and departures. Our framework is more general than the methods prevalent in the literature in that it is applicable to a broad range of problems characterized by a variety of action polytopes and generic arrival and departure processes. In Chapter 4, we consider the problem of A-B testing when the impact of the treatment is marred by a large number of covariates. Randomization can be highly inefficient in such settings, and thus we consider the problem of optimally allocating test subjects to either treatment with a view to maximizing the efficiency of our estimate of the treatment effect. Our main contribution is a tractable algorithm for this problem in the online setting, where subjects arrive, and must be assigned, sequentially. We characterize the value of optimized allocations relative to randomized allocations and show that this value grows large as the number of covariates grows. In particular, we show that there is a lot to be gained from ‘optimizing’ the process of A-B testing relative to the simple randomized trials that are the mainstay of A-B testing in the ‘big data’ regime of modern e-commerce applications, where the number of covariates is often comparable to the number of experimental trials.Dynamic programming, Stochastic approximation, Stochastic control theory, Markov processes, Operations researchnb2490BusinessThesesEssays on Cloud Pricing and Causal Inference
https://academiccommons.columbia.edu/catalog/ac:200357
Kilcioglu, Cinar10.7916/D8R78F9QThu, 15 Jun 2017 15:04:37 +0000In this thesis, we study economics and operations of cloud computing, and we propose new matching methods in observational studies that enable us to estimate the effect of green building practices on market rents.
In the first part, we study a stylized revenue maximization problem for a provider of cloud computing services, where the service provider (SP) operates an infinite capacity system in a market with heterogeneous customers with respect to their valuation and congestion sensitivity. The SP offers two service options: one with guaranteed service availability, and one where users bid for resource availability and only the "winning" bids at any point in time get access to the service. We show that even though capacity is unlimited, in several settings, depending on the relation between valuation and congestion sensitivity, the revenue maximizing service provider will choose to make the spot service option stochastically unavailable. This form of intentional service degradation is optimal in settings where user valuation per unit time increases sub-linearly with respect to their congestion sensitivity (i.e., their disutility per unit time when the service is unavailable) -- this is a form of "damaged goods." We provide some data evidence based on the analysis of price traces from the biggest cloud service provider, Amazon Web Services.
In the second part, we study the competition on price and quality in cloud computing. The public "infrastructure as a service" cloud market possesses unique features that make it difficult to predict long-run economic behavior. On the one hand, major providers buy their hardware from the same manufacturers, operate in similar locations and offer a similar menu of products. On the other hand, the competitors use different proprietary "fabric" to manage virtualization, resource allocation and data transfer. The menus offered by each provider involve a discrete number of choices (virtual machine sizes) and allow providers to locate in different parts of the price-quality space. We document this differentiation empirically by running benchmarking tests. This allows us to calibrate a model of firm technology. Firm technology is an input into our theoretical model of price-quality competition. The monopoly case highlights the importance of competition in blocking "bad equilibrium" where performance is intentionally slowed down or options are unduly limited. In duopoly, price competition is fierce, but prices do not converge to the same level because of price-quality differentiation. The model helps explain market trends, such the healthy operating profit margin recently reported by Amazon Web Services. Our empirically calibrated model helps not only explain price cutting behavior but also how providers can manage a profit despite predictions that the market "should be" totally commoditized.
The backbone of cloud computing is datacenters, whose energy consumption is enormous. In the past years, there has been an extensive effort on making the datacenters more energy efficient. Similarly, buildings are in the process going "green" as they have a major impact on the environment through excessive use of resources. In the last part of this thesis, we revisit a previous study about the economics of environmentally sustainable buildings and estimate the effect of green building practices on market rents. For this, we use new matching methods that take advantage of the clustered structure of the buildings data. We propose a general framework for matching in observational studies and specific matching methods within this framework that simultaneously achieve three goals: (i) maximize the information content of a matched sample (and, in some cases, also minimize the variance of a difference-in-means effect estimator); (ii) form the matches using a flexible matching structure (such as a one-to-many/many-to-one structure); and (iii) directly attain covariate balance as specified ---before matching--- by the investigator. To our knowledge, existing matching methods are only able to achieve, at most, two of these goals simultaneously. Also, unlike most matching methods, the proposed methods do not require estimation of the propensity score or other dimensionality reduction techniques, although with the proposed methods these can be used as additional balancing covariates in the context of (iii). Using these matching methods, we find that green buildings have 3.3% higher rental rates per square foot than otherwise similar buildings without green ratings ---a moderately larger effect than the one previously found.Data processing service centers--Energy conservation, Cloud computing, Industrial management, Industrial engineering, Operations researchck2560BusinessThesesEssays on Approximation Algorithms for Robust Linear Optimization Problems
https://academiccommons.columbia.edu/catalog/ac:202072
Lu, Brian Yin10.7916/D8ZG6SGMThu, 15 Jun 2017 15:04:37 +0000Solving optimization problems under uncertainty has been an important topic since the appearance of mathematical optimization in the mid 19th century. George Dantzig’s 1955 paper, “Linear Programming under Uncertainty” is considered one of the ten most inﬂuential papers in Management Science [26]. The methodology introduced in Dantzig’s paper is named stochastic programming, since it assumes an underlying probability distribution of the uncertain input parameters. However, stochastic programming suffers from the “curse of dimensionality”, and knowing the exact distribution of the input parameter may not be realistic. On the other hand, robust optimization models the uncertainty using a deterministic uncertainty set. The goal is to optimize the worst-case scenario from the uncertainty set. In recent years, many studies in robust optimization have been conducted and we refer the reader to Ben-Tal and Nemirovski [4–6], El Ghaoui and Lebret [19], Bertsimas and
Sim [15, 16], Goldfarb and Iyengar [23], Bertsimas et al. [8] for a review of robust optimization. Computing an optimal adjustable (or dynamic) solution to a robust optimization problem is generally hard. This motivates us to study the hardness of approximation of the problem and provide efﬁcient approximation algorithms. In this dissertation, we consider adjustable robust linear optimization problems with packing and covering formulations and their approximation algorithms. In particular, we study the performances of static solution and afﬁne solution as approximations for the adjustable robust problem.
Chapter 2 and 3 consider two-stage adjustable robust linear packing problem with uncertain second-stage constraint coefﬁcients. For general convex, compact and down-monotone uncertainty sets, the problem is often intractable since it requires to compute a solution for all possible realizations of uncertain parameters [22]. In particular, for a fairly general class of uncertainty sets, we show that the two-stage adjustable robust problem is NP-hard to approximate within a factor that is better than Ω(logn), where n is the number of columns of the uncertain coefﬁcient matrix. On the other hand, a static solution is a single (here and now) solution that is feasible for all possible realizations of the uncertain parameters and can be computed efﬁciently. We study the performance of static solutions an approximation for the adjustable robust problem and relate its optimality to a transformation of the uncertain set. With this transformation, we show that for a fairly general class of uncertainty sets, static solution is optimal for the adjustable robust problem. This is surprising since the static solution is widely perceived as highly conservative. Moreover, when the static solution is not optimal, we provide an instance-based tight approximation bound that is related to a measure of non-convexity of the transformation of the uncertain set. We also show that for two-stage problems, our bound is at least as good (and in many case signiﬁcantly better) as the bound given by the symmetry of the uncertainty set [11, 12]. Moreover, our results can be generalized to the case where the objective coefﬁcients and right-hand-side are also uncertainty.
In Chapter 3, we focus on the two-stage problems with a family of column-wise and constraint-wise uncertainty sets where any constraint describing the set involves entries of only a single column or a single row. This is a fairly general class of uncertainty sets to model constraint coefﬁcient uncertainty. Moreover, it is the family of uncertainty sets that gives the previous hardness result. On the positive side, we show that a static solution is an
O(\log n · min(\log \Gamma, \log(m+n))-approximation for the two-stage adjustable robust problem where m and n denote the numbers of rows and columns of the constraint matrix and \Gamma is the maximum possible ratio of upper bounds of the uncertain constraint coefﬁcients. Therefore, for constant \Gamma, surprisingly the performance bound for static solutions matches
the hardness of approximation for the adjustable problem. Furthermore, in general the static solution provides nearly the best efﬁcient approximation for the two-stage adjustable robust problem.
In Chapter 4, we extend our result in Chapter 2 to a multi-stage adjustable robust linear optimization problem. In particular, we consider the case where the choice of the uncertain constraint coefﬁcient matrix for each stage is independent of the others. In real world applications, decision problems are often of multiple stages and a iterative implementation of two-stage solution may result in a suboptimal solution for multi-stage problem. We consider the static solution for the adjustable robust problem and the transformation of the uncertainty sets introduced in Chapter 2. We show that the static solution is optimal for the adjustable robust problem when the transformation of the uncertainty set for each stage is convex.
Chapters 5 considers a two-stage adjustable robust linear covering problem with uncertain right-hand-side parameter. As mentioned earlier, such problems are often intractable due to astronomically many extreme points of the uncertainty set. We introduce a new approximation framework where we consider a “simple” set that is “close” to the original uncertainty set. Moreover, the adjustable robust problem can be solved efﬁciently over the extended set. We show that the approximation bound is related to a geometric factor that represents the Banach-Mazur distance between the two sets. Using this framework, we provide approximation bounds that are better than the bounds given by an afﬁne policy in [7] for a large class of interesting uncertainty sets. For instance, we provide an approximation solution that gives a m^{1/4}-approximation for the two-stage adjustable robust problem with hypersphere uncertainty set, while the afﬁne policy has an approximation ratio of O(\sqrt{m}).
Moreover, our bound for general p-norm ball is m^{\frac{p-1}{p^2}} as opposed to m^{1/p} as given by an affine policy.Mathematical optimization, Uncertainty (Information theory), Robust optimization, Approximation algorithms, Operations researchyl2662Industrial Engineering and Operations ResearchThesesPricing Models in the Presence of Informational and Social Externalities
https://academiccommons.columbia.edu/catalog/ac:200421
Crapis, Davide10.7916/D8222TWPThu, 15 Jun 2017 15:04:37 +0000This thesis studies three game theoretic models of pricing, in which a seller is interested in optimally pricing and allocating her product or service to a market of agents, in order to maximize her revenue. These markets feature a large number of self-interested agents, who are generally heterogeneous with respect to some payoff relevant feature, e.g., willingness to pay when agents are consumers or private cost when agents are firms. Agents strategically interact with one another, and their actions affect other agents' payoffs, either directly through social influence or competition, or indirectly through a review system. The seller has demand uncertainty and strives to optimize expected discounted revenues. I use either a mean-field approximation or a continuum of agents assumption to reduce the complexity of the problems and provide crisp characterizations of system aggregates and equilibrium policies.
Chapter 2 considers the problem of an information provider who sells information products, such as demand forecasts, to a market of firms that compete with one another in a downstream market. We propose a general model that subsumes both price and quantity competition as special cases. We characterize the optimal selling strategy and find that it depends on both mode and intensity of competition. Several important extensions to heterogeneous production costs, information quality discrimination, and information leakage through aggregate actions are studied.
Chapter 3 considers the problem of optimally extracting a stream of revenues from a sequence of consumers, who have heterogeneous willingness to pay and uncertainty about the quality of the product being sold. Using an intuitive maximum likelihood procedure, we characterize the solution of consumers' quality estimation problem. Then, we use a mean-field approximation to characterize the transient dynamics of quality estimates and demand. These allow us to simplify and solve the monopolist's problem, and ultimately provide a characterization of her optimal pricing policy.
Chapter 4 considers the problem of a seller who is interested in dynamically pricing her product when consumers' utility is influenced by the mass of consumers that have purchased in the past. Two scenarios are studied, one in which the monopolist has commitment power and one in which she does not. We characterize the optimal selling strategy under both scenarios and derive comparisons on equilibrium prices and demands. Our main result is a characterization of the value of price commitment as a function of the social influence level in the market.Consumer behavior, Game theory--Economic aspects, Competition, Prices, Operations researchdc2792BusinessThesesOn the Trade-oﬀs between Modeling Power and Algorithmic Complexity
https://academiccommons.columbia.edu/catalog/ac:202359
Ye, Chun10.7916/D87W6CDCThu, 15 Jun 2017 15:04:10 +0000Mathematical modeling is a central component of operations research. Most of the academic research in our field focuses on developing algorithmic tools for solving various mathematical problems arising from our models. However, our procedure for selecting the best model to use in any particular application is ad hoc. This dissertation seeks to rigorously quantify the trade-offs between various design criteria in model construction through a series of case studies. The hope is that a better understanding of the pros and cons of different models (for the same application) can guide and improve the model selection process.
In this dissertation, we focus on two broad types of trade-offs. The first type arises naturally in mechanism or market design, a discipline that focuses on developing optimization models for complex multi-agent systems. Such systems may require satisfying multiple objectives that are potentially in conflict with one another. Hence, finding a solution that simultaneously satisfies several design requirements is challenging. The second type addresses the dynamics between model complexity and computational tractability in the context of approximation algorithms for some discrete optimization problems. The need to study this type of trade-offs is motivated by certain industry problems where the goal is to obtain the best solution within a reasonable time frame. Hence, being able to quantify and compare the degree of sub-optimality of the solution obtained under different models is helpful. Chapters 2-5 of the dissertation focus on trade-offs of the first type and Chapters 6-7 the second type.Operations research, Multiagent systems, Mathematical optimization, Approximation algorithms, Mathematical modelscy2214Industrial Engineering and Operations ResearchThesesMicrostructure Analysis of Dynamic Markets: Limit Order Books and Dynamic Matching Markets
https://academiccommons.columbia.edu/catalog/ac:194157
Zheng, Hua10.7916/D82V2FXHThu, 15 Jun 2017 15:02:38 +0000This thesis is concerned with addressing operational issues in two types of dynamic markets where queueing plays an important role: limit order books (financial industry), and dynamic matching markets (residential real estate).
We first study the smart order routing decisions of investors in fragmented limit order book markets and the implications on the market dynamics. In modern equity markets, participants have a choice of many exchanges at which to trade. Exchanges typically operate as electronic limit order books operating under a “price-time” priority rule and, in turn, can be modeled as multi-class FIFO queueing systems. A market with multiple exchanges can be thought as a decentralized, parallel queueing system. Heterogeneous traders that submit limit orders select the exchange to place their orders by trading off delays until their order may fill against financial considerations. Simultaneously, traders that submit market orders select the exchange to direct their orders. These market orders trigger instantaneous service completions of queued limit orders. Taking into account the effect of investors’ order routing decisions, we find that the equilibrium of this decentralized market exhibits a state space collapse property. The predicted dimension reduction is the result of high-frequency order routing decisions that essentially couple the dynamics across exchanges. Analyzing a TAQ dataset for a sample of stocks over a one month period, we find empirical support for the predicted state space collapse.
In the second part of this thesis, we model an electronic limit order book as a multi-class queueing system under fluid dynamics, and formulate and solve a problem of limit and market order placement to optimally buy a block of shares over a short, predetermined time horizon. Using the structure of the optimal execution policy, we identify microstructure variables that affect trading costs over short time horizons and propose a resulting microstructure-based model of market impact costs. We use a proprietary data set to estimate this cost model, and highlight its insightful structure and increased accuracy over conventional (macroscopic) market impact models that estimate the cost of a trade based on its normalized size but disregarding measurements of limit order book variables.
In the third part of this thesis, we study the residential real estate markets as dynamic matching systems with an emphasis on their microstructure. We propose a stylized microstructure model and analyze the market dynamics and its equilibrium under the simplifying approximation where buyers and sellers use linear bidding strategies. We motivate and characterize this near closed-form approximation of the market equilibrium, and show that it is asymptotically accurate. We also provide numerical evidence in support of this approximation. Then with the gained tractability, we characterize steady-state properties such as market depth, price dispersion, and anticipated delays in selling or buying a unit. We characterize congestion and matching patterns for sellers and buyers, taking into account market dynamics, heterogeneity, and supply and demand imbalance manifested in the competition among buyers and sellers. Furthermore, we show the effects of market primitives with comparative statics results.Residential real estate, Stock exchanges, Queuing theory, Congestion pricing, Finance, Public, Operations researchhz2215BusinessThesesResource Allocation in Wireless Networks: Theory and Applications
https://academiccommons.columbia.edu/catalog/ac:202119
Marasevic, Jelena Rajko10.7916/D85T3KP0Thu, 15 Jun 2017 15:01:50 +0000Limited wireless resources, such as spectrum and maximum power, give rise to various resource allocation problems that are interesting both from theoretical and application viewpoints. While the problems in some of the wireless networking applications are amenable to general resource allocation methods, others require a more specialized approach suited to their unique structural characteristics. We study both types of the problems in this thesis.
We start with a general problem of alpha-fair packing, namely, the problem of maximizing sum_j {w_j f_α(x_j)}, where w_j > 0, ∀j, and (i) f_α(x_j)=ln(x_j), if α = 1, (ii) f_α(x_j)= {x_j^(1-α)}/{1-α}, if α ≠ 1,α > 0, subject to positive linear constraints of the form Ax ≤ b, x ≥ 0, where A and b are non-negative. This problem has broad applications within and outside wireless networking. We present a distributed algorithm for general alpha that converges to an epsilon-approximate solution in time (number of distributed iterations) that has an inverse polynomial dependence on the approximation parameter epsilon and poly-logarithmic dependence on the problem size. This is the first distributed algorithm for weighted alpha-fair packing with poly-logarithmic convergence in the input size. We also obtain structural results that characterize alpha-fair allocations as the value of alpha is varied. These results deepen our understanding of fairness guarantees in alpha-fair packing allocations, and also provide insights into the behavior of alpha-fair allocations in the asymptotic cases when alpha tends to zero, one, and infinity.
With these general tools on hand, we consider an application in wireless networks where fairness is of paramount importance: rate allocation and routing in energy-harvesting networks. We discuss the importance of fairness in such networks and cases where our results on alpha-fair packing apply. We then turn our focus to rate allocation in energy harvesting networks with highly variable energy sources and that are used for applications such as monitoring and tracking. In such networks, it is essential to guarantee fairness over both the network nodes and the time slots and to be as fair as possible -- in particular, to require max-min fairness. We first develop an algorithm that obtains a max-min fair rate assignment for any routing that is specified at the input. Then, we consider the problem of determining a "good'' routing. We consider various routing types and either provide polynomial-time algorithms for finding such routings or prove that the problems are NP-hard. Our results reveal an interesting trade-off between the complexities of computation and implementation. The results can also be applied to other related fairness problems.
The second part of the thesis is devoted to the study of resource allocation problems that require a specialized approach. The problems we focus on arise in wireless networks employing full-duplex communication -- the simultaneous transmission and reception on the same frequency channel. Our primary goal is to understand the benefits and complexities tied to using this novel wireless technology through the study of resource (power, time, and channel) allocation problems. Towards that goal, we introduce a new realistic model of a compact (e.g., smartphone) full-duplex receiver and demonstrate its accuracy via measurements. First, we focus on the resource allocation problems with the objective of maximizing the sum of uplink and downlink rates, possibly over multiple orthogonal channels. For the single-channel case, we quantify the rate improvement as a function of the remaining self-interference and signal-to-noise ratios and provide structural results that characterize the sum of uplink and downlink rates on a full-duplex channel. Building on these results, we consider the multi-channel case and develop a polynomial time algorithm which is nearly optimal in practice under very mild restrictions. To reduce the running time, we develop an efficient nearly-optimal algorithm under the high SINR approximation.
Then, we study the achievable capacity regions of full-duplex links in the single- and multi-channel cases. We present analytical results that characterize the uplink and downlink capacity region and efficient algorithms for computing rate pairs at the region's boundary. We also provide near-optimal and heuristic algorithms that "convexify'' the capacity region when it is not convex. The convexified region corresponds to a combination of a few full-duplex rates (i.e., to time sharing between different operation modes). The analytical results provide insights into the properties of the full-duplex capacity region and are essential for future development of fair resource allocation and scheduling algorithms in Wi-Fi and cellular networks incorporating full-duplex.Resource allocation, Electrical engineering, Energy harvesting, Wireless sensor networks, Operations research, Computer sciencejrm2207Electrical EngineeringThesesApplied Inventory Management: New Approaches to Age-Old Problems
https://academiccommons.columbia.edu/catalog/ac:194202
Daniel Guetta, Charles Raphael10.7916/D84M94B1Wed, 14 Jun 2017 21:12:18 +0000Supply chain management is one of the fundamental topics in the field of operations research, and a vast literature exists on the subject. Many recent developments in the field are rapidly narrowing the gap between the systems handled in the literature and the real-life problems companies need to solve on a day-to-day basis. However, there are certain features often observed in real-world systems that elude even these most recent developments. In this thesis, we consider a number of these features, and propose some new heuristics together with methodologies to evaluate their performance.
In Chapter 2, we consider a general two-echelon distribution system consisting of a depot and multiple sales outlets which face random demands for a given item. The replenishment process consists of two stages: the depot procures the item from an outside supplier, while the retailers' inventories are replenished by shipments from the depot. Both of the replenishment stages are associated with a given facility-specific leadtime. The depot as well as the retailers face a limited inventory capacity. We propose a heuristic for this class of dynamic programming models to obtain an upper bound on optimal costs, together with a new approach to generate lower bounds based on Lagrangian relaxation. We report on an extensive numerical study with close to 14,000 instances which evaluates the accuracy of the lower bound and the optimality gap of the various heuristic policies. Our study reveals that our policy performs exceedingly well almost across the entire parameter spectrum.
In Chapter 3, we extend the model above to deal with distribution systems involving several items. In this setting, two interdependencies can arise between items that considerably complicate the problem. First, shared storage capacity at each of the retail outlets results in a trade-off between items; ordering more of one item means less space is available for another. Second, economies of scope can occur in the order costs if several items can be ordered from a single supplier, incurring only one fixed cost. To our knowledge, our approach is the first that has been proposed to handle such complex, multi-echelon, multi-item systems. We propose a heuristic for this class of dynamic programming models, to obtain an upper bound on optimal costs, together with an approach to generate lower bounds. We report on an extensive numerical study with close to 1,200 instances that reveals our heuristic performs excellently across the entire parameter spectrum. In Chapter 4, we consider a periodic-review stochastic inventory control system consisting of a single retailer which faces random demands for a given item, and in which demand forecasts are dynamically updated (for example, new information observed in one period may affect our beliefs about demand distributions in future periods). Replenishment orders are subject to fixed and variable costs. A number of heuristics exist to deal with such systems, but to our knowledge, no general approach exists to find lower bounds on optimal costs therein. We develop a general approach for finding lower bounds on the cost of such systems using an information relaxation. We test our approach in a model with advance demand information, and obtain good lower bounds over a range of problem parameters.
Finally, in Appendix A, we begin to tackle the problem of using these methods in real supply chain systems. We were able to obtain data from a luxury goods manufacturer to inspire our study. Unfortunately, the methods we developed in earlier chapters were not directly applicable to these data. Instead, we developed some alternate heuristic methods, and we considered statistical techniques that might be used to obtain the parameters required for these heuristics from the data available.Industrial management, Inventory control, Inventory control--Data processing, Inventory control--Evaluation, Inventory control--Management, Retail trade--Inventory control, Business logistics, Operations research, Industrial engineeringBusinessThesesDemand Learning in Two Operations Models
https://academiccommons.columbia.edu/catalog/ac:194987
Han, Yunru10.7916/D8CZ3704Wed, 14 Jun 2017 21:12:02 +0000The rapid advance of information technologies largely facilitated firms' data-driven decision making. Particularly, in operations management practices, firms could continuously collect information to refine their demand knowledge, and integrate this process into their relevant operational decisions, e.g. pricing, inventory, and market entry, known as demand learning. Demand learning in complex business systems is often tangled with complex strategic interactions, thus requiring a deep understanding of how it affects the strategic relationship among players in various business setups. This thesis aims to contribute to the demand-learning literature by studying the strategic interactions in two different business relationships, one vertical and the other horizontal.
First, I consider the interactions between a retailer and a supplier in a supply chain subject to demand censorship (i.e. unobservable lost sales) when the retailer is engaging in demand learning through dynamic inventory experimentation. I study the supplier's optimal wholesale prices when the retailer is in three different situations, and find that the retailer and the supply chain may actually benefit from either myopia or censorship in contrast to the existing results, due to the supplier's different collaborative or exploitative responses to the retailer's "willingness to learn". I also identify that, with demand censorship, the collaborative behavior between the players for information acquisition may improve the system's performance.
Second, I study an online retail platform's learning process and entry policies as well as the independent seller's pricing distortion behavior to slow down this process, motivated by Amazon.com's unique dual role as both a marketplace and a merchant that allows it to use the transaction data generated by its third-party sellers to decide if to sell the same product itself. I developed a Bayesian statistical model for the platform's demand learning, proposed two types of heuristic entry policies for the platform owner. The model predicts a pattern of price distortion, and describes the product offering choices made by the independent seller. These could potentially serve as testable results for empirical studies.Decision making, Pricing, Business logistics, Electronic commerce, Supply and demand, Operations management, Operations research, Industrial management, EconomicsBusinessThesesPerfect Simulation and Deployment Strategies for Detection
https://academiccommons.columbia.edu/catalog/ac:189976
Wallwater, Aya10.7916/D8X066JBWed, 14 Jun 2017 19:56:13 +0000This dissertation contains two parts. The first part provides the first algorithm that, under minimal assumptions, allows to simulate the stationary waiting-time sequence of a single-server queue backwards in time, jointly with the input processes of the queue
(inter-arrival and service times).
The single-server queue is useful in applications of DCFTP (Dominated Coupling From The Past), which is a well known protocol for simulation without bias from steady-state distributions. Our algorithm terminates in finite time assuming only finite mean of the
inter-arrival and service times. In order to simulate the single-server queue in stationarity until the first idle period in finite expected termination time we require the existence of finite variance. This requirement is also necessary for such idle time (which is a natural
coalescence time in DCFTP applications) to have finite mean. Thus, in this sense, our algorithm is applicable under minimal assumptions.
The second part studies the behavior of diffusion processes in a random environment.
We consider an adversary that moves in a given domain and our goal is to produce an optimal strategy to detect and neutralize him by a given deadline. We assume that the target's dynamics follows a diffusion process whose parameters are informed by available intelligence information. We will dedicate one chapter to the rigorous formulation of the detection problem, an introduction of several frameworks that can be considered when applying our methods, and a discussion on the challenges of finding the analytical optimal solution. Then, in the following chapter, we will present our main result, a large deviation behavior of the adversary's survival probability under a given strategy. This result will be later give rise to asymptotically efficient Monte Carlo algorithms.Operations researchaw2589Industrial Engineering and Operations ResearchThesesSmart Grid Risk Management
https://academiccommons.columbia.edu/catalog/ac:188373
Abad Lopez, Carlos Adrian10.7916/D8028QR9Wed, 14 Jun 2017 19:55:13 +0000Current electricity infrastructure is being stressed from several directions -- high demand, unreliable supply, extreme weather conditions, accidents, among others. Infrastructure planners have, traditionally, focused on only the cost of the system; today, resilience and sustainability are increasingly becoming more important. In this dissertation, we develop computational tools for efficiently managing electricity resources to help create a more reliable and sustainable electrical grid. The tools we present in this work will help electric utilities coordinate demand to allow the smooth and large scale integration of renewable sources of energy into traditional grids, as well as provide infrastructure planners and operators in developing countries a framework for making informed planning and control decisions in the presence of uncertainty.
Demand-side management is considered as the most viable solution for maintaining grid stability as generation from intermittent renewable sources increases. Demand-side management, particularly demand response (DR) programs that attempt to alter the energy consumption of customers either by using price-based incentives or up-front power interruption contracts, is more cost-effective and sustainable in addressing short-term supply-demand imbalances when compared with the alternative that involves increasing fossil fuel-based fast spinning reserves. An essential step in compensating participating customers and benchmarking the effectiveness of DR programs is to be able to independently detect the load reduction from observed meter data. Electric utilities implementing automated DR programs through direct load control switches are
also interested in detecting the reduction in demand to efficiently pinpoint non-functioning devices to reduce maintenance costs. We develop sparse optimization methods for detecting a small change in the demand for electricity of a customer in response to a price change or signal from the utility, dynamic learning methods for scheduling the maintenance of direct load control switches whose operating state is not directly observable and can only be inferred from the metered electricity consumption, and machine learning methods for accurately forecasting the load of hundreds of thousands of residential, commercial and industrial customers. These algorithms have been implemented in the software system provided by AutoGrid, Inc., and this system has helped several utilities in the Pacific Northwest, Oklahoma, California and Texas, provide more reliable power to their customers at significantly reduced prices.
Providing power to widely spread out communities in developing countries using the conventional power grid is not economically feasible. The most attractive alternative source of affordable energy for these communities is solar micro-grids. We discuss risk-aware robust methods to optimally size and operate solar micro-grids in the presence of uncertain demand and uncertain renewable generation. These algorithms help system operators to increase their revenue while making their systems more resilient to inclement weather conditions.Operations research, Power resourcesca2446Industrial Engineering and Operations ResearchThesesHeavy Tails and Instabilities in Large-Scale Systems with Failures
https://academiccommons.columbia.edu/catalog/ac:189193
Skiani, Evangelia10.7916/D8MW2GKRWed, 14 Jun 2017 19:54:59 +0000Modern engineering systems, e.g., wireless communication networks, distributed computing systems, etc., are characterized by high variability and susceptibility to failures. Failure recovery is required to guarantee the successful operation of these systems. One straight- forward and widely used mechanism is to restart the interrupted jobs from the beginning after a failure occurs. In network design, retransmissions are the primary building blocks of the network architecture that guarantee data delivery in the presence of channel failures. Retransmissions have recently been identified as a new origin of power laws in modern information networks. In particular, it was discovered that retransmissions give rise to long tails (delays) and possibly zero throughput. To this end, we investigate the impact of the ‘retransmission phenomenon’ on the performance of failure prone systems and propose adaptive solutions to address emerging instabilities.
The preceding finding of power law phenomena due to retransmissions holds under the assumption that data sizes have infinite support. In practice, however, data sizes are upper bounded 0 ≤ L ≤ b, e.g., WaveLAN’s maximum transfer unit is 1500 bytes, YouTube videos are of limited duration, e-mail attachments cannot exceed 10MB, etc. To this end, we first provide a uniform characterization of the entire body of the distribution of the number of retransmissions, which can be represented as a product of a power law and the Gamma distribution. This rigorous approximation clearly demonstrates the transition from power law distributions in the main body to exponential tails. Furthermore, the results highlight the importance of wisely determining the size of data fragments in order to accommodate the performance needs in these systems as well as provide the appropriate tools for this fragmentation.
Second, we extend the analysis to the practically important case of correlated channels using modulated processes, e.g., Markov modulated, to capture the underlying dependencies. Our study shows that the tails of the retransmission and delay distributions are asymptotically insensitive to the channel correlations and are determined by the state that generates the lightest tail in the independent channel case. This insight is beneficial both for capacity planning and channel modeling since the independent model is sufficient and the correlation details do not matter. However, the preceding finding may be overly optimistic when the best state is atypical, since the effects of ‘bad’ states may still downgrade the performance.
Third, we examine the effects of scheduling policies in queueing systems with failures and restarts. Fair sharing, e.g., processor sharing (PS), is a widely accepted approach to resource allocation among multiple users. We revisit the well-studied M/G/1 PS queue with a new focus on server failures and restarts. Interestingly, we discover a new phenomenon showing that PS-based scheduling induces complete instability in the presence of retransmissions, regardless of how low the traffic load may be. This novel phenomenon occurs even when the job sizes are bounded/fragmented, e.g., deterministic. This work demonstrates that scheduling one job at a time, such as first-come-first-serve, achieves a larger stability region and should be preferred in these systems.
Last, we delve into the area of distributed computing and study the effects of commonly used mechanisms, i.e., restarts, fragmentation, replication, especially in cloud computing services. We evaluate the efficiency of these techniques under different assumptions on the data streams and discuss the corresponding optimization problem. These findings are useful for optimal resource allocation and fault tolerance in rapidly developing computing networks. In addition to networking and distributed computing systems, the aforementioned results improve our understanding of failure recovery management in large manufacturing and service systems, e.g., call centers. Scalable solutions to this problem increase in significance as these systems continuously grow in scale and complexity. The new phenomena and the techniques developed herein provide new insights in the areas of parallel computing, probability and statistics, as well as financial engineering.Electrical engineering, Operations researches3009Electrical EngineeringThesesTwo Essays in Financial Engineering
https://academiccommons.columbia.edu/catalog/ac:189643
Yang, Linan10.7916/D8K35T1MWed, 14 Jun 2017 19:54:52 +0000This dissertation consists of two parts. In the first part, we investigate the potential impact of wrong-way risk on calculating credit valuation adjustment (CVA) of a derivatives portfolio. A credit valuation adjustment (CVA) is an adjustment applied to the value of a derivative contract or a portfolio of derivatives to account for counterparty credit risk. Measuring CVA requires combining models of market and credit risk. Wrong-way risk refers to the possibility that a counterparty's likelihood of default increases with the market value of the exposure. We develop a method for bounding wrong-way risk, holding fixed marginal models for market and credit risk and varying the dependence between them. Given simulated paths of the two models, we solve a linear program to find the worst-case CVA resulting from wrong-way risk. We analyze properties of the solution and prove convergence of the estimated bound as the number of paths increases. The worst case can be overly pessimistic, so we extend the procedure for a tempered CVA by penalizing the deviation of the joint model of market and credit risk from a reference model. By varying the penalty for deviations, we can sweep out the full range of possible CVA values for different degrees of wrong-way risk. Here, too, we prove convergence of the estimate of the tempered CVA and the joint distribution that attains it. Our method addresses an important source of model risk in counterparty risk measurement. In the second part, we study investors' trading behavior in a model of realization utility. We assume that investors' trading decisions are driven not only by the utility of consumption and terminal wealth, but also by the utility burst from realizing a gain or a loss. More precisely, we consider a dynamic trading problem in which an investor decides when to purchase and sell a stock to maximize her wealth utility and realization utility with her reference points adapting to the stock's gain and loss asymmetrically. We study, both theoretically and numerically, the optimal trading strategies and asset pricing implications of two types of agents: adaptive agents, who realize prospectively the reference point adaptation in the future, and naive agents, who fail to do so. We find that an adaptive agent sells the stock more frequently when the stock is at a gain than a naive agent does, and that the adaptive agent asks for a higher risk premium for the stock than the naive agent does in equilibrium. Moreover, compared to a non-adaptive agent whose reference point does not change with the stock's gain and loss, both the adaptive and naive agents sell the stock less frequently, and the naive agent requires the same risk premium as the non-adaptive agent does.Operations research, Financely2220Industrial Engineering and Operations ResearchThesesRanking Algorithms on Directed Configuration Networks
https://academiccommons.columbia.edu/catalog/ac:189652
Chen, Ningyuan10.7916/D8J38RX8Wed, 14 Jun 2017 19:54:45 +0000In recent decades, complex real-world networks, such as social networks, the World Wide Web, financial networks, etc., have become a popular subject for both researchers and practitioners. This is largely due to the advances in computing power and big-data analytics. A key issue of analyzing these networks is the centrality of nodes. Ranking algorithms are designed to achieve the goal, e.g., Google's PageRank. We analyze the asymptotic distribution of the rank of a randomly chosen node, computed by a family of ranking algorithms on a random graph, including PageRank, when the size of the network grows to infinity.
We propose a configuration model generating the structure of a directed graph given in- and out-degree distributions of the nodes. The algorithm guarantees the generated graph to be simple (without self-loops and multiple edges in the same direction) for a broad spectrum of degree distributions, including power-law distributions. Power-law degree distribution is referred to as scale-free property and observed in many real-world networks. On the random graph G_n=(V_n,E_n) generated by the configuration model, we study the distribution of the ranks, which solves
R_i=∑ _{j: (j,i) ∈ E_n} (C_jR_j +Q_i)
for all node i, some weight C_i and personalization value Q_i.
We show that as the size of the graph n → ∞, the rank of a randomly chosen node converges weakly to the endogenous solution of the
R =^D ∑ _{i=1}^N (C_iR_i + Q),
where (Q, N, {C_i}) is a random vector and {R_i} are i.i.d. copies of R, independent of (Q, N,{C_i}). This main result is divided into three steps. First, we show that the rank of a randomly chosen node can be approximated by applying the ranking algorithm on the graph for finite iterations. Second, by coupling the graph to a branching tree that is governed by the empirical size-biased distribution, we approximate the finite iteration of the ranking algorithm by the root node of the branching tree. Finally, we prove that the rank of the root of the branching tree converges to that of a limiting weighted branching process, which is independent of n and solves the stochastic fixed-point equation. Our result formalizes the well-known heuristics, that a network often locally possesses a tree-like structure. We conduct a numerical example showing that the approximation is very accurate for English Wikipedia pages (over 5 million).
To draw a sample from the endogenous solution of the stochastic fixed-point equation, one can run linear branching recursions on a weighted branching process. We provide an iterative simulation algorithm based on bootstrap. Compared to the naive Monte Carlo, our algorithm reduces the complexity from exponential to linear in the number of recursions. We show that as the bootstrap sample size tends to infinity, the sample drawn according to our algorithm converges to the target distribution in the Kantorovich-Rubinstein distance and the estimator is consistent.Operations research, Computer sciencenc2462Industrial Engineering and Operations ResearchThesesDecision Making with Coupled Learning: Applications in Inventory Management and Auctions
https://academiccommons.columbia.edu/catalog/ac:188400
Chaneton, Juan Manuel10.7916/D86M3632Wed, 14 Jun 2017 19:54:38 +0000Operational decisions can be complicated by the presence of uncertainty. In many cases, there exist means to reduce uncertainty, though these may come at a cost. Decision makers then face the dilemma of acting based on current, incomplete information versus investing in trying to minimize uncertainty. Understanding the impact of this trade-off on decisions and performance is the central topic of this thesis.
When attempting to construct probabilistic models based on data, operational decisions often affect the amount and quality of data that is collected. This introduces an exploration-exploitation trade-off between decisions and information collection. Much of the literature has sought to understand how operational decisions should be modified to incorporate this trade-off. While studying two well-known operational problems, we ask an even more basic question: does the exploration-exploitation trade-off matter in the first place? In the first two parts of this thesis we focus on this question in the context of the newsvendor problem and sequential auctions with incomplete private information.
We first analyze the well-studied stationary multi-period newsvendor problem, in which a retailer sells perishable items and unmet demand is lost and unobserved. This latter limitation, referred to as demand censoring, is what introduces the exploration-exploitation trade-off in this problem. We focus on two questions: i.) what is the value of accounting for the exploration-exploitation trade-off; and, ii.) what is the cost imposed by having access only to sales data as opposed to underlying demand samples? Quite remarkably, we show that, for a broad family of tractable cases, there is essentially no exploration-exploitation trade-off; i.e., there is almost no value of accounting for the impact of decisions on information collection. Moreover, we establish that losses due to demand censoring (as compared to having full access to demand samples) are limited, but these are of higher order than those due to ignoring the exploration-exploitation trade-off. In other words, efforts aimed at improving information collection concerning lost sales are more valuable than analytic or computational efforts to pin down the optimal policy in the presence of censoring.
In the second part of this thesis we examine the problem of an agent bidding on a sequence of repeated auctions for an item. The agent does not fully know his own valuation of the object and he can only collect information if he wins an auction. This coupling introduces the exploration-exploitation trade-off in this problem. We study the value of accounting for information collection on decisions and find that: i.) in general the exploration-exploitation trade-off cannot be ignored (that is, in some cases ignoring exploration can substantially affect rewards), but ii.) for a broad class of instances, ignoring exploration can indeed produce nearly optimal results. We characterize this class through a set of conditions on the problem primitives, and we demonstrate with examples that these are satisfied for common settings found in the literature.
In the third part of this thesis we study the impact of uncertainty in the context of inventory record inaccuracies in inventory management systems. Record inaccuracies, mismatches between physical and recorded inventory, are frequently encountered in practice and can markedly affect revenues. Most of the literature is devoted to analyzing the cost-benefit relationship between investing in means to reduce inaccuracies and accounting for them in operational decisions. We focus on the less explored approach of using available data to reduce the uncertainty in inventory. In practice, collecting Point Of Sale (POS) data is substantially simpler than collecting stock information. We propose a model in which inventory is regarded as a virtually unobservable quantity and POS data is used to infer its state over time. Additionally, our method also works as an effective estimator of censored demand in the presence of inaccurate records. We test our methodology with extensive numerical experiments based on both simulated and actual retailing data. The results show that it is remarkably effective in inferring unobservable past statistics and predicting future stock status, even in the presence of severe data misspecifications.Operations research, Businessjmc2274BusinessThesesStochastic Networks: Modeling, Simulation Design and Risk Control
https://academiccommons.columbia.edu/catalog/ac:189655
Li, Juan10.7916/D88P5ZV3Wed, 14 Jun 2017 19:53:54 +0000This dissertation studies stochastic network problems that arise in various areas with important industrial applications. Due to uncertainty of both external and internal variables, these networks are exposed to the risk of failure with certain probability, which, in many cases, is very small. It is thus desirable to develop efficient simulation algorithms to study the stability of these networks and provide guidance for risk control.
Chapter 2 models equilibrium allocations in a distribution network as the solution of a linear program (LP) which minimizes the cost of unserved demands across nodes in the network. Assuming that the demands are random (following a jointly Gaussian law), we study the probability that the optimal cost exceeds a large threshold, which is a rare event. Our contribution is the development of importance sampling and conditional Monte Carlo algorithms for estimating this probability. We establish the asymptotic efficiency of our algorithms and also present numerical results that demonstrate the strong performance of our algorithms.
Chapter 3 studies an insurance-reinsurance network model that deals with default contagion risks with a particular aim of capturing cascading effects at the time of defaults. We capture these effects by finding an equilibrium allocation of settlements that can be found as the unique optimal solution of an optimization problem. We are able to obtain an asymptotic description of the most likely ways in which the default of a specific group of participants can occur, by solving a multidimensional Knapsack integer programming problem. We also propose a class of strongly efficient Monte Carlo estimators for computing the expected loss of the network conditioned on the failure of a specific set of companies.
Chapter 4 discusses control schemes for maintaining low failure probability of a transmission system power line. We construct a stochastic differential equation to describe the temperature evolution in a line subject to stochastic exogenous factors such as ambient temperature, and present a solution to the resulting stochastic heat equation. A number of control algorithms designed to limit the probability that a line exceeds its critical temperature are provided.Operations research, Engineering, Financejl3035Industrial Engineering and Operations ResearchThesesAsymptotic Analysis of Service Systems with Congestion-Sensitive Customers
https://academiccommons.columbia.edu/catalog/ac:193897
Yao, John Jia-Hao10.7916/D86D5SR9Wed, 14 Jun 2017 19:52:00 +0000Many systems in services, manufacturing, and technology, feature users or customers sharing a limited number of resources, and which suffer some form of congestion when the number of users exceeds the number of resources. In such settings, queueing models are a common tool for describing the dynamics of the system and quantifying the congestion that results from the aggregated effects of individuals joining and leaving the system. Additionally, the customers themselves may be sensitive to congestion and react to the performance of the system, creating feedback and interaction between individual customer behavior and aggregate system dynamics.This dissertation focuses on the modeling and performance of service systems with congestion-sensitive customers using large-scale asymptotic analyses of queueing models. This work extends the theoretical literature on congestion-sensitive customers in queues in the settings of service differentiation and observational learning and abandonment. Chapter 2 considers the problem of a service provider facing a heterogeneous market of customers who differ based on their value for service and delay sensitivity. The service provider seeks to find the revenue maximizing level of service differentiation (offering different price-delay combinations). We show that the optimal policy places the system in heavy traffic, but at substantially different levels of congestion depending on the degree of service differentiation. Moreover, in a differentiated offering, the level of congestion will vary substantially between service classes. Chapter 3 presents a new model of customer abandonment in which congestion-sensitive customers observe the queue length, but do not know the service rate. Instead, they join the queue and observe their progress in order to estimate their wait times and make abandonment decisions. We show that an overloaded queue with observational learning and abandonment stabilizes at a queue length whose scale depends on the tail of the service time distribution. Methodologically, our asymptotic approach leverages stochastic limit theory to provide simple and intuitive results for optimizing or characterizing system performance. In particular, we use the analysis of deterministic fluid-type queues to provide a first-order characterization of the stochastic system dynamics, which is demonstrated by the convergence of the stochastic system to the fluid model. This also allows us to crisply illustrate and quantify the relative contributions of system or customer characteristics to overall system performance.Operations research, Queuing theory, Dynamics, Consumer satisfactionjjy2107BusinessThesesOptimization methods for power grid reliability
https://academiccommons.columbia.edu/catalog/ac:194133
Harnett, Sean R.10.7916/D8MS3SM8Wed, 14 Jun 2017 19:50:00 +0000This dissertation focuses on two specific problems related to the reliability of the modern power grid. The first part investigates the economic dispatch problem with uncertain power sources. The classic economic dispatch problem seeks generator power output levels that meet demand most efficiently; we add risk-awareness to this by explicitly modeling the uncertainty of intermittent power sources using chance-constrained optimization and incorporating the chance constraints into the standard optimal power flow framework. The result is a dispatch of power which is substantially more robust to random fluctuations with only a small increase in economic cost. Furthermore, it uses an algorithm which is only moderately slower than the conventional practice.
The second part investigates “the power grid attack problem”: aiming to maximize disruption to the grid, how should an attacker distribute a budget of “damage” across the power lines? We formulate it as a continuous problem, which bypasses the combinatorial explosion of a discrete formulation and allows for interesting attacks containing lines that are only partially damaged rather than completely removed. The result of our solution to the attack problem can provide helpful information to grid planners seeking to improve the resilience of the power grid to outages and disturbances. Both parts of this dissertation include extensive experimental results on a number of cases, including many realistic large-scale instances.Electric power systems--Reliability--Mathematical models, Electric power distribution--Reliability, Mathematical optimization, Electric power systems--Reliability, Mathematics, Operations researchsrh2144Applied Physics and Applied MathematicsThesesThe Theory of Systemic Risk
https://academiccommons.columbia.edu/catalog/ac:178176
Chen, Chen10.7916/D8W37TWCMon, 12 Jun 2017 17:40:26 +0000Systemic risk is an issue of great concern in modern financial markets as well as, more broadly, in the management of complex business and engineering systems. It refers to the risk of collapse of an entire complex system, as a result of the actions taken by the individual component entities or agents that comprise the system. We investigate the topic of systemic risk from the perspectives of measurement, structural sources, and risk factors. In particular, we propose an axiomatic framework for the measurement and management of systemic risk based on the simultaneous analysis of outcomes across agents in the system and over scenarios of nature. Our framework defines a broad class of systemic risk measures that accommodate a rich set of regulatory preferences. This general class of systemic risk measures captures many specific measures of systemic risk that have recently been proposed as special cases, and highlights their implicit assumptions. Moreover, the systemic risk measures that satisfy our conditions yield decentralized decompositions, i.e., the systemic risk can be decomposed into risk due to individual agents. Furthermore, one can associate a shadow price for systemic risk to each agent that correctly accounts for the externalities of the agent's individual decision-making on the entire system. Also, we provide a structural model for a financial network consisting of a set of firms holding common assets. In the model, endogenous asset prices are captured by the marketing clearing condition when the economy is in equilibrium. The key ingredients in the financial market that are captured in this model include the general portfolio choice flexibility of firms given posted asset prices and economic states, and the mark-to-market wealth of firms. The price sensitivity can be analyzed, where we characterize the key features of financial holding networks that minimize systemic risk, as a function of overall leverage. Finally, we propose a framework to estimate risk measures based on risk factors. By introducing a form of factor-separable risk measures, the acceptance set of the original risk measure connects to the acceptance sets of the factor-separable risk measures. We demonstrate that the tight bounds for factor-separable coherent risk measures can be explicitly constructed.Operations researchcc3136Industrial Engineering and Operations ResearchThesesMethods for Pricing Pre-Earnings Equity Options and Leveraged ETF Options
https://academiccommons.columbia.edu/catalog/ac:186986
Santoli, Marco10.7916/D86Q1W99Mon, 12 Jun 2017 17:39:26 +0000In this thesis, we present several analytical and numerical methods for two financial engineering problems: 1) accounting for the impact of an earnings announcement on the price and implied volatility of the associated equity options, and 2) analyzing the price dynamics of leveraged exchange-traded funds (LETFs) and valuation of LETF options. Our pricing models capture the main characteristics of these options, along with jumps and stochastic volatility in the underlying asset. We illustrate our results through numerical implementation and calibration using market data.
In the first part, we model the pricing of equity options around an earnings announcement (EA). Empirical studies have shown that an earnings announcement can lead to an immediate price shock to the company stock. Since many companies also have options written on their stocks, the option prices should reflect the uncertain price impact of an upcoming EA before expiration. To represent the shock due to earnings, we incorporate a random jump on the announcement date in the dynamics of the stock price. We consider different distributions of the scheduled earnings jump as well as different underlying stock price dynamics before and after the EA date. Our main contributions include analytical option pricing formulas when the underlying stock price follows the Kou model along with a double-exponential or Gaussian EA jump on the announcement date. Furthermore, we derive analytic bounds and asymptotics for the pre-EA implied volatility under various models. The calibration results demonstrate adequate fit of the entire implied volatility surface prior to an announcement. The comparison of the risk-neutral distribution of the EA jump to its historical counterpart is also discussed. Moreover, we discuss the valuation and exercise strategy of pre-EA American options, and present an analytical approximation and numerical results.
The second part focuses on the analysis of LETFs. We start by providing a quantitative risk analysis of LETFs with an emphasis on the impact of leverage ratios and investment horizons. Given an investment horizon, different leverage ratios imply different levels of risk. Therefore, the idea of an {admissible range of leverage ratios} is introduced. For an admissible leverage ratio, the associated LETF satisfies a given risk constraint based on, for example, the value-at-risk (VaR) and conditional VaR. Moreover, we discuss the concept of {admissible risk horizon} so that the investor can control risk exposure by selecting an appropriate holding period. The intra-horizon risk is calculated, showing that higher leverage can significantly increase the probability of an LETF value hitting a lower level. This leads us to evaluate a stop-loss/take-profit strategy for LETFs and determine the optimal take-profit given a stop-loss risk constraint. In addition, the impact of volatility exposure on the returns of different LETF portfolios is investigated.
In the last chapter, we study the pricing of options written on LETFs. Since LETFs on the same reference index share the same source of risk, it is important to consider a consistent pricing methodology of these options. In addition, LETFs can theoretically experience a loss greater than 100\%. In practice, some LETF providers design the fund so that the daily returns are capped both downward and upward. We incorporate these features and model the reference index by a stochastic volatility model with jumps. An efficient numerical algorithm based on transform methods to value options under this model is presented. We illustrate the accuracy of our pricing algorithm by comparing it to existing methods. Calibration using empirical option data shows the impact of leverage ratio on the implied volatility. Our method is extended to price American-style LETF options.Finance, Operations researchIndustrial Engineering and Operations ResearchThesesOptimal Multiple Stopping Approach to Mean Reversion Trading
https://academiccommons.columbia.edu/catalog/ac:186941
Li, Xin10.7916/D88K781SMon, 12 Jun 2017 17:39:20 +0000This thesis studies the optimal timing of trades under mean-reverting price dynamics subject to fixed transaction costs. We first formulate an optimal double stopping problem whereby a speculative investor can choose when to enter and subsequently exit the market. The investor's value functions and optimal timing strategies are derived when prices are driven by an Ornstein-Uhlenbeck (OU), exponential OU, or Cox-Ingersoll-Ross (CIR) process. Moreover, we analyze a related optimal switching problem that involves an infinite sequence of trades. In addition to solving for the value functions and optimal switching strategies, we identify the conditions under which the double stopping and switching problems admit the same optimal entry and/or exit timing strategies. A number of extensions are also considered, such as incorporating a stop-loss constraint, or a minimum holding period under the OU model.
A typical solution approach for optimal stopping problems is to study the associated free boundary problems or variational inequalities (VIs). For the double optimal stopping problem, we apply a probabilistic methodology and rigorously derive the optimal price intervals for market entry and exit. A key step of our approach involves a transformation, which in turn allows us to characterize the value function as the smallest concave majorant of the reward function in the transformed coordinate. In contrast to the variational inequality approach, this approach directly constructs the value function as well as the optimal entry and exit regions, without a priori conjecturing a candidate value function or timing strategy. Having solved the optimal double stopping problem, we then apply our results to deduce a similar solution structure for the optimal switching problem. We also verify that our value functions solve the associated VIs.
Among our results, we find that under OU or CIR price dynamics, the optimal stopping problems admit the typical buy-low-sell-high strategies. However, when the prices are driven by an exponential OU process, the investor generally enters when the price is low, but may find it optimal to wait if the current price is sufficiently close to zero. In other words, the continuation (waiting) region for entry is disconnected. A similar phenomenon is observed in the OU model with stop-loss constraint. Indeed, the entry region is again characterized by a bounded price interval that lies strictly above the stop-loss level. As for the exit timing, a higher stop-loss level always implies a lower optimal take-profit level. In all three models, numerical results are provided to illustrate the dependence of timing strategies on model parameters.Operations research, Financexl2426Industrial Engineering and Operations ResearchThesesEssays on Inventory Management and Conjoint Analysis
https://academiccommons.columbia.edu/catalog/ac:181257
Chen, Yupeng10.7916/D8GX49BDMon, 12 Jun 2017 17:38:22 +0000With recent theoretic and algorithmic advancements, modern optimization methodologies have seen a substantial expansion of modeling power, being applied to solve challenging problems in impressively diverse areas. This dissertation aims to extend the modeling frontier of optimization methodologies in two exciting fields inventory management and conjoint analysis. Although
the three essays concern distinct applications using different optimization methodologies, they
share a unifying theme, which is to develop intuitive models using advanced optimization techniques to solve problems of practical relevance. The first essay (Chapter 2) applies robust optimization to solve a single installation inventory model with non stationary uncertain demand. A classical problem in operations research, the inventory management model could become very challenging to analyze when lost sales dynamics, non zero fixed ordering cost, and positive lead time are introduced. In this essay, we propose a robust cycle based control policy based on an innovative decomposition idea to solve a family of variants of this model. The policy is simple, flexible, easily implementable and numerical experiments suggest that the policy has very promising empirical performance.The policy can be used both when the excess demand is backlogged as well as when it is lost; with non zero fixed ordering cost, and also when lead time is non zero. The policy decisions are computed by solving a collection of linear programs even when there is a positive fixed ordering cost. The policy also extends in a very simple manner to the joint pricing and inventory control problem. The second essay (Chapter 3) applies sparse machine learning to model multimodal continuous heterogeneity in conjoint analysis. Consumers' heterogeneous preferences can often be represented using a multimodal continuous heterogeneity (MCH) distribution. One interpretation of MCH is that the consumer population consists of a few distinct segments, each of which contains a heterogeneous sub population. Modeling of MCH raises considerable challenges as both across and within segment heterogeneity need to be accounted for. In this essay, we propose an innovative sparse learning approach for modeling MCH and apply it to conjoint analysis where adequate modeling of consumer heterogeneity is critical. The sparse learning approach models MCH via a two-stage divide and conquer framework, in which we first decompose the consumer population by recovering a set of candidate segmentations using structured sparsity modeling, and then use each candidate segmentation to develop a set of individual level representations of MCH. We select the optimal individual level representation of MCH and the corresponding optimal candidate segmentation using cross-validation. Two notable features of our approach are that it accommodates both across and within segment heterogeneity and endogenously imposes an adequate amount of shrinkage to recover the individual level partworths. We empirically validate the performance of the sparse learning approach using extensive simulation experiments and two empirical conjoint data sets. The third essay (Chapter 4) applies dynamic discrete choice models to investigate the impact of return policies on consumers' product purchase and return behavior. Return policies have been ubiquitous in the marketplace, allowing consumers to use and evaluate a product before fully committing to purchase. Despite the clear practical relevance of return policies, however, few studies have provided empirical assessments of how consumers' purchase and return decisions respond to the return policies facing them. In this essay, we propose to model consumers' purchase and return decisions using a dynamic discrete choice model with forward looking and Bayesian learning. More specifically, we postulate that consumers' purchase and return decisions are optimal solutions for some underlying dynamic expected utility maximization problem in which consumers learn their true evaluations of products via usage in a Bayesian manner and make purchase and return decisions to maximize their expected present value of utility, and return policies impact consumers' purchase and return decisions by entering the dynamic expected utility maximization problem as constraints. Our proposed model provides a behaviorally plausible approach to examine the impact of return policies on consumers' purchase and return behavior.Operations researchyc2561Industrial Engineering and Operations ResearchThesesA Family of Latent Variable Convex Relaxations for IBM Model 2
https://academiccommons.columbia.edu/catalog/ac:207865
Simion, Andrei10.7916/D8NV9HBNMon, 12 Jun 2017 17:36:27 +0000Introduced in 1993, the IBM translation models were the first generation Statistical Machine Translation systems. For the IBM Models, only IBM Model 1 is a convex optimization problem, meaning that we can initialize all its probabilistic parameters to uniform values and subsequently converge to a good solution via Expectation Maximization (EM). In this thesis we discuss a mechanism to generate an infinite supply of nontrivial convex relaxations for IBM Model 2 and detail an Exponentiated Subgradient algorithm to solve them. We also detail some interesting relaxations that admit and easy EM algorithm that does not require the tuning of a learning rate. Based on the geometric mean of two variables, this last set of convex models can be seamlessly integrated into the open-source GIZA++ word-alignment library. Finally, we also show other applications of the method, including a more powerful strictly convex IBM Model 1, and a convex HMM surrogate that improves on the performance of the previous convex IBM Model 2 variants.Translating machines, Operations research, Machine translating, Computer science, Artificial intelligenceIndustrial Engineering and Operations ResearchThesesExcluding Induced Paths: Graph Structure and Coloring
https://academiccommons.columbia.edu/catalog/ac:186521
Maceli, Peter Lawson10.7916/D8WW7GK4Mon, 12 Jun 2017 17:35:55 +0000An induced subgraph of a given graph is any graph which can be obtained by successively deleting vertices, possible none. In this thesis, we present several new structural and algorithmic results on a number of different classes of graphs which are closed under taking induced subgraphs.
The first result of this thesis is related to a conjecture of Hayward and Nastos on the structure of graphs with no induced four-edge path or four-edge antipath. They conjectured that every such graph which is both prime and perfect is either a split graph or contains a certain useful arrangement of simplicial and antisimplicial vertices. We give a counterexample to their conjecture, and prove a slightly weaker version. This is joint work with Maria Chudnovsky, and first appeared in Journal of Graph Theory.
The second result of this thesis is a decomposition theorem for the class of all graphs with no induced four-edge path or four-edge antipath. We show that every such graph can be obtained from pentagons and split graphs by repeated application of complementation, substitution, and split graph unification. Split graph unification is a new graph operation we introduced, which is a generalization of substitution and involves "gluing" two graphs along a common induced split graph. This is a combination of joint work with Maria Chudnovsky and Irena Penev, together with later work of Louis Esperet, Laetitia Lemoine and Frederic Maffray, and first appeared in.
The third result of this thesis is related to the problem of determining the complexity of coloring graphs which do not contain some fixed induced subgraph. We show that three-coloring graphs with no induced six-edge path or triangle can be done in polynomial-time. This is joint work with Maria Chudnovsky and Mingxian Zhong, and first appeared in. Working together with Flavia Bonomo, Oliver Schaudt, and Maya Stein, we have since simplified and extended this result.Operations research, Mathematics, Computer scienceplm2109Industrial Engineering and Operations ResearchThesesStochastic Approximation Algorithms in the Estimation of Quasi-Stationary Distribution of Finite and General State Space Markov Chains
https://academiccommons.columbia.edu/catalog/ac:177124
Zheng, Shuheng10.7916/D89C6VM9Thu, 08 Jun 2017 20:24:23 +0000This thesis studies stochastic approximation algorithms for estimating the quasi-stationary distribution of Markov chains. Existing numerical linear algebra methods and probabilistic methods might be computationally demanding and intractable in large state spaces. We take our motivation from a heuristic described in the physics literature and use the stochastic approximation framework to analyze and extend it.
The thesis begins by looking at the finite dimensional setting. The finite dimensional quasi-stationary estimation algorithm was proposed in the Physics literature by [#latestoliveira, #oliveiradickman1, #dickman], however no proof was given there and it was not recognized as a stochastic approximation algorithm. This and related schemes were analyzed in the context of urn problems and the consistency of the estimator is shown there [#aldous1988two, #pemantle, #athreya]. The rate of convergence is studied by [#athreya] in special cases only. The first chapter provides a different proof of the algorithm's consistency and establishes a rate of convergence in more generality than [#athreya]. It is discovered that the rate of convergence is only fast when a certain restrictive eigenvalue condition is satisfied. Using the tool of iterate averaging, the algorithm can be modified and we can eliminate the eigenvalue condition.
The thesis then moves onto the general state space discrete-time Markov chain setting. In this setting, the stochastic approximation framework does not have a strong theory in the current literature, so several of the convergence results have to be adapted because the iterates of our algorithm are measure-valued The chapter formulates the quasi-stationary estimation algorithm in this setting. Then, we extend the ODE method of [#kushner2003stochastic] and proves the consistency of algorithm. Through the proof, several non-restrictive conditions required for convergence of the algorithm are discovered.
Finally, the thesis tests the algorithm by running some numerical experiments. The examples are designed to test the algorithm in various edge cases. The algorithm is also empirically compared against the Fleming-Viot method.Operations researchIndustrial Engineering and Operations ResearchThesesHigh-Dimensional Portfolio Management: Taxes, Execution and Information Relaxations
https://academiccommons.columbia.edu/catalog/ac:185815
Wang, Chun10.7916/D8M043JJThu, 08 Jun 2017 20:22:56 +0000Portfolio management has always been a key topic in finance research area. While many researchers have studied portfolio management problems, most of the work to date assumes trading is frictionless. This dissertation presents our investigation of the optimal trading policies and efforts of applying duality method based on information relaxations to portfolio problems where the investor manages multiple securities and confronts trading frictions, in particular capital gain taxes and execution cost.
In Chapter 2, we consider dynamic asset allocation problems where the investor is required to pay capital gains taxes on her investment gains. This is a very challenging problem because the tax to be paid whenever a security is sold depends on the tax basis, i.e. the price(s) at which the security was originally purchased. This feature results in high-dimensional and path-dependent problems which cannot be solved exactly except in the case of very stylized problems with just one or two securities and relatively few time periods. The asset allocation problem with taxes has several variations depending on: (i) whether we use the exact or average tax-basis and (ii) whether we allow the full use of losses (FUL) or the limited use of losses (LUL). We consider all of these variations in this chapter but focus mainly on the exact and average-cost tax-basis LUL cases since these problems are the most realistic and generally the most challenging. We develop several sub-optimal trading policies for these problems and use duality techniques based on information relaxations to assess their performances. Our numerical experiments consider problems with as many as 20 securities and 20 time periods. The principal contribution of this chapter is in demonstrating that much larger problems can now be tackled through the use of sophisticated optimization techniques and duality methods based on information-relaxations. We show in fact that the dual formulation of exact tax-basis problems are much easier to solve than the corresponding primal problems. Indeed, we can easily solve dual problem instances where the number of securities and time periods is much larger than 20. We also note, however, that while the average tax-basis problem is relatively easier to solve in general, its corresponding dual problem instances are non-convex and more difficult to solve. We therefore propose an approach for the average tax-basis dual problem that enables valid dual bounds to still be obtained.
In Chapter 3, we consider a portfolio execution problem where a possibly risk-averse agent needs to trade a fixed number of shares in multiple stocks over a short time horizon. Our price dynamics can capture linear but stochastic temporary and permanent price impacts as well as stochastic volatility. In general it's not possible to solve even numerically for the optimal policy in this model, however, and so we must instead search for good sub-optimal policies. Our principal policy is a variant of an open-loop feedback control (OLFC) policy and we show how the corresponding OLFC value function may be used to construct good primal and dual bounds on the optimal value function. The dual bound is constructed using the recently developed duality methods based on information relaxations. One of the contributions of this chapter is the identification of sufficient conditions to guarantee convexity, and hence tractability, of the associated dual problem instances. That said, we do not claim that the only plausible models are those where all dual problem instances are convex. We also show that it is straightforward to include a non-linear temporary price impact as well as return predictability in our model. We demonstrate numerically that good dual bounds can be computed quickly even when nested Monte-Carlo simulations are required to estimate the so-called dual penalties. These results suggest that the dual methodology can be applied in many models where closed-form expressions for the dual penalties cannot be computed.
In Chapter 4, we apply duality methods based on information relaxations to dynamic zero-sum games. We show these methods can easily be used to construct dual lower and upper bounds for the optimal value of these games. In particular, these bounds can be used to evaluate sub-optimal policies for zero-sum games when calculating the optimal policies and game value is intractable.Operations research, FinanceIndustrial Engineering and Operations ResearchThesesNew Quantitative Approaches to Asset Selection and Portfolio Construction
https://academiccommons.columbia.edu/catalog/ac:175867
Song, Irene10.7916/D83N21JVThu, 08 Jun 2017 20:21:08 +0000Since the publication of Markowitz's landmark paper "Portfolio Selection" in 1952, portfolio construction has evolved into a disciplined and personalized process. In this process, security selection and portfolio optimization constitute key steps for making investment decisions across a collection of assets. The use of quantitative algorithms and models in these steps has become a widely-accepted investment practice by modern investors. This dissertation is devoted to exploring and developing those quantitative algorithms and models.
In the first part of the dissertation, we present two efficiency-based approaches to security selection: (i) a quantitative stock selection strategy based on operational efficiency and (ii) a quantitative currency selection strategy based on macroeconomic efficiency. In developing the efficiency-based stock selection strategy, we exploit a potential positive link between firm's operational efficiency and its stock performance. By means of data envelopment analysis (DEA), a non-parametric approach to productive efficiency analysis, we quantify firm's operational efficiency into a single score representing a consolidated measure of financial ratios. The financial ratios integrated into an efficiency score are selected on the basis of their predictive power for the firm's future operating performance using the LASSO (least absolute shrinkage and selection operator)-based variable selection method. The computed efficiency scores are directly used for identifying stocks worthy of investment. The basic idea behind the proposed stock selection strategy is that as efficient firms are presumed to be more profitable than inefficient firms, higher returns are expected from their stocks. This idea is tested in a contextual and empirical setting provided by the U.S. Information Technology (IT) sector. Our empirical findings confirm that there is a strong positive relationship between firm's operational efficiency and its stock performance, and further establish that firm's operational efficiency has significant explanatory power in describing the cross-sectional variations of stock returns. We moreover offer an economic argument that posits operational efficiency as a systematic risk factor and the most likely source of excess returns of investing in efficient firms.
The efficiency-based currency selection strategy is developed in a similar way; i.e. currencies are selected based on a certain efficiency metric. An exchange rate has long been regarded as a reliable barometer of the state of the economy and the measure of international competitiveness of countries. While strong and appreciating currencies correspond to productive and efficient economies, weak and depreciating currencies correspond to slowing down and less efficient economies. This study hence develops a currency selection strategy that utilizes macroeconomic efficiency of countries measured based on a widely-accepted relationship between exchange rates and macroeconomic variables. For quantifying macroeconomic efficiency of countries, we first establish a multilateral framework using effective exchange rates and trade-weighted macroeconomic variables. This framework is used for transforming the three representative bilateral structural exchange rate models: the flexible price monetary model, the sticky price monetary model, and the sticky price asset model, into their multilateral counterparts. We then translate these multilateral models into DEA models, which yield an efficiency score representing an aggregate measure of macroeconomic variables. Consistent with the stock selection strategy, the resulting efficiency scores are used for identifying currencies worthy of investment. We evaluate our currency selection strategy against appropriate market and strategic benchmarks using historical data. Our empirical results confirm that currencies of efficient countries have stronger performance than those of inefficient countries, and further suggest that compared to the exchange rate models based on standard regression analysis, our models based on DEA improve on the predictability of the future performance of currencies.
In the first part of the dissertation, we also develop a data-driven variable selection method for DEA based on the group LASSO. This method extends the LASSO-based variable selection method used for specifying a DEA model for estimating firm's operational efficiency. In our proposed method, we derive a special constrained version of the group LASSO with the loss function suited for variable selection in DEA models and solve it by a new tailored algorithm based on the alternating direction method of multipliers (ADMM). We conduct a thorough evaluation of the proposed method against two widely-used variable selection methods: the efficiency contribution measure (ECM) method and the regression-based (RB) test, in the DEA literature using Monte Carlo simulations. The simulation results show that our method provides more favorable performance compared with its benchmarks.
In the second part of the dissertation, we propose a generalized risk budgeting (GRB) approach to portfolio construction. In a GRB portfolio, assets are grouped into possibly overlapping subsets, and each subset is allocated a risk budget that has been pre-specified by the investor. Minimum variance, risk parity and risk budgeting portfolios are all special instances of a GRB portfolio. The GRB portfolio optimization problem is to find a GRB portfolio with an optimal risk-return profile where risk is measured using any positively homogeneous risk measure. When the subsets form a partition, the assets all have identical returns and we restrict ourselves to long-only portfolios, then the GRB problem can in fact be solved as a convex optimization problem. In general, however, the GRB problem is a constrained non-convex problem, for which we propose two solution approaches. The first approach uses a semidefinite programming (SDP) relaxation to obtain an (upper) bound on the optimal objective function value. In the second approach we develop a numerical algorithm that integrates augmented Lagrangian and Markov chain Monte Carlo (MCMC) methods in order to find a point in the vicinity of a very good local optimum. This point is then supplied to a standard non-linear optimization routine with the goal of finding this local optimum. It should be emphasized that the merit of this second approach is in its generic nature: in particular, it provides a starting-point strategy for any non-linear optimization algorithms.Operations researchIndustrial Engineering and Operations ResearchThesesGraph Structure and Coloring
https://academiccommons.columbia.edu/catalog/ac:175631
Plumettaz, Matthieu10.7916/D87M0637Thu, 08 Jun 2017 20:19:17 +0000We denote by G=(V,E) a graph with vertex set V and edge set E. A graph G is claw-free if no vertex of G has three pairwise nonadjacent neighbours. Claw-free graphs are a natural generalization of line graphs. This thesis answers several questions about claw-free graphs and line graphs.
In 1988, Chvatal and Sbihi proved a decomposition theorem for claw-free perfect graphs. They showed that claw-free perfect graphs either have a clique-cutset or come from two basic classes of graphs called elementary and peculiar graphs. In 1999, Maffray and Reed successfully described how elementary graphs can be built using line graphs of bipartite graphs and local augmentation. However gluing two claw-free perfect graphs on a clique does not necessarily produce claw-free graphs. The first result of this thesis is a complete structural description of claw-free perfect graphs. We also give a construction for all perfect circular interval graphs. This is joint work with Chudnovsky.
Erdos and Lovasz conjectured in 1968 that for every graph G and all integers s,t≥ 2 such that s+t-1=χ(G) > ω(G), there exists a partition (S,T) of the vertex set of G such that ω(G|S)≥ s and χ(G|T)≥ t. This conjecture is known in the graph theory community as the Erdos-Lovasz Tihany Conjecture. For general graphs, the only settled cases of the conjecture are when s and t are small. Recently, the conjecture was proved for a few special classes of graphs: graphs with stability number 2, line graphs and quasi-line graphs. The second part of this thesis considers the conjecture for claw-free graphs and presents some progresses on it. This is joint work with Chudnovsky and Fradkin.
Reed's ω, ∆, χ conjecture proposes that every graph satisfies χ≤ ⎡½ (Δ+1+ω)⎤ ; it is known to hold for all claw-free graphs. The third part of this thesis considers a local strengthening of this conjecture. We prove the local strengthening for line graphs, then note that previous results immediately tell us that the local strengthening holds for all quasi-line graphs. Our proofs lead to polytime algorithms for constructing colorings that achieve our bounds: The complexity are O(n²) for line graphs and O(n³m²) for quasi-line graphs. For line graphs, this is faster than the best known algorithm for constructing a coloring that achieves the bound of Reed's original conjecture. This is joint work with Chudnovsky, King and Seymour.Operations researchmp2761Industrial Engineering and Operations ResearchThesesEssays on Infrastructure Design and Planning for Clean Energy Systems
https://academiccommons.columbia.edu/catalog/ac:176967
Kocaman, Ayse10.7916/D8JW8C2FThu, 08 Jun 2017 20:15:36 +0000The International Energy Agency estimates that the number of people who do not have access to electricity is nearly 1.3 billion and a billion more have only unreliable and intermittent supply. Moreover, current supply for electricity generation mostly relies on fossil fuels, which are finite and one of the greatest threats to the environment. Rising population growth rates, depleting fuel sources, environmental issues and economic developments have increased the need for mathematical optimization to provide a formal framework that enables systematic and clear decision-making in energy operations. This thesis through its methodologies and algorithms enable tools for energy generation, transmission and distribution system design and help policy makers make cost assessments in energy infrastructure planning rapidly and accurately.
In Chapter 2, we focus on local-level power distribution systems planning for rural electrification using techniques from combinatorial optimization. We describe a heuristic algorithm that provides a quick solution for the partial electrification problem where the distribution network can only connect a pre-specified number of households with low voltage lines. The algorithm demonstrates the effect of household settlement patterns on the electrification cost. We also describe the first heuristic algorithm that selects the locations and service areas of transformers without requiring candidate solutions and simultaneously builds a two-level grid network in a green-field setting. The algorithms are applied to real world rural settings in Africa, where household locations digitized from satellite imagery are prescribed.
In Chapter 3 and 4, we focus on power generation and transmission using clean energy sources. Here, we imagine a country in the future where hydro and solar are the dominant sources and fossil fuels are only available in minimal form. We discuss the problem of modeling hydro and solar energy production and allocation, including long-term investments and storage, capturing the stochastic nature of hourly supply and demand data. We mathematically model two hybrid energy generation and allocation systems where time variability of energy sources and demand is balanced using the water stored in the reservoirs. In Chapter 3, we use conventional hydro power stations (incoming stream flows are stored in large dams and water release is deferred until it is needed) and in Chapter 4, we use pumped hydro stations (water is pumped from lower reservoir to upper reservoir during periods of low demand to be released for generation when demand is high). Aim of the models is to determine optimal sizing of infrastructure needed to match demand and supply in a most reliable and cost effective way.
An innovative contribution of this work is the establishment of a new perspective to energy modeling by including fine-grained sources of uncertainty such as stream flow and solar radiations in hourly level as well as spatial location of supply and demand and transmission network in national level. In addition, we compare the conventional and the pumped hydro power systems in terms of reliability and cost efficiency and quantitatively show the improvement provided by including pumped hydro storage. The model will be presented with a case study of India and helps to answer whether solar energy in addition to hydro power potential in Himalaya Mountains would be enough to meet growing electricity demand if fossil fuels could be almost completely phased out from electricity generation.Environmental engineering, Operations research, Power resourcesask2170Earth and Environmental EngineeringThesesData-driven Decisions in Service Systems
https://academiccommons.columbia.edu/catalog/ac:175604
Kim, Song-Hee10.7916/D8D798KHThu, 08 Jun 2017 20:15:15 +0000This thesis makes contributions to help provide data-driven (or evidence-based) decision support to service systems, especially hospitals. Three selected topics are presented.
First, we discuss how Little's Law, which relates average limits and expected values of stationary distributions, can be applied to service systems data that are collected over a finite time interval. To make inferences based on the indirect estimator of average waiting times, we propose methods for estimating confidence intervals and for adjusting estimates to reduce bias. We show our new methods are effective using simulations and data from a US bank call center.
Second, we address important issues that need to be taken into account when testing whether real arrival data can be modeled by nonhomogeneous Poisson processes (NHPPs). We apply our method to data from a US bank call center and a hospital emergency department and demonstrate that their arrivals come from NHPPs.
Lastly, we discuss an approach to standardize the Intensive Care Unit admission process, which currently lacks a well-defined criteria. Using data from nearly 200,000 hospitalizations, we discuss how we can quantify the impact of Intensive Care Unit admission on individual patient's clinical outcomes. We then use this quantified impact and a stylized model to discuss optimal admission policies. We use simulation to compare the performance of our proposed optimal policies to the current admission policy, and show that the gain can be significant.Operations researchsk3116Industrial Engineering and Operations ResearchThesesConvex Optimization Algorithms and Recovery Theories for Sparse Models in Machine Learning
https://academiccommons.columbia.edu/catalog/ac:175385
Huang, Bo10.7916/D8VM49DMThu, 08 Jun 2017 20:14:25 +0000Sparse modeling is a rapidly developing topic that arises frequently in areas such as machine learning, data analysis and signal processing. One important application of sparse modeling is the recovery of a high-dimensional object from relatively low number of noisy observations, which is the main focuses of the Compressed Sensing, Matrix Completion(MC) and Robust Principal Component Analysis (RPCA) . However, the power of sparse models is hampered by the unprecedented size of the data that has become more and more available in practice. Therefore, it has become increasingly important to better harnessing the convex optimization techniques to take advantage of any underlying "sparsity" structure in problems of extremely large size.
This thesis focuses on two main aspects of sparse modeling. From the modeling perspective, it extends convex programming formulations for matrix completion and robust principal component analysis problems to the case of tensors, and derives theoretical guarantees for exact tensor recovery under a framework of strongly convex programming. On the optimization side, an efficient first-order algorithm with the optimal convergence rate has been proposed and studied for a wide range of problems of linearly constraint sparse modeling problems.Mathematics, Statistics, Operations researchIndustrial Engineering and Operations ResearchThesesOn the Kidney Exchange Problem and Online Minimum Energy Scheduling
https://academiccommons.columbia.edu/catalog/ac:175610
Herrera Humphries, Tulia10.7916/D8125QSXThu, 08 Jun 2017 20:14:25 +0000The allocation and management of scarce resources are of central importance in the design
of policies to improve social well-being. This dissertation consists of three essays; the first
two deals with the problem of allocating kidneys and the third one on power management
in computing devices.
Kidney exchange programs are an attractive alternative for patients who need a kidney
transplant and who have a willing, but medically incompatible, donor. A registry that keeps
track of such patient-donor pairs can nd matches through exchanges amongst such pairs.
This results in a quicker transplant for the patients involved, and equally importantly, keeps
such patients from the long wait list of patients without an intended donor. As of March
2014, there were at least 99,000 candidates waiting for a kidney transplant in the U.S.
However, in 2013 only 16,893 transplants were conducted. This imbalance between supply
and demand among other factors, has driven the development of multiple kidney exchange
programs in the U.S. and the subsequent development of matching mechanisms to run the
programs.
In the first essay we consider a matching problem arising in kidney exchanges between
hospitals. Focusing on the case of two hospitals, we construct a strategy-proof matching
mechanism that is guaranteed to return a matching that is at least 3/4 the size of a maximum cardinality
matching. It is known that no better performance is possible if one focuses on
mechanisms that return a maximal matching, and so our mechanism is best possible within
this natural class of mechanisms. For path-cycle graphs we construct a mechanism that
returns a matching that is at least 4/5 the size of max-cardinality matching. This mechanism
does not necessarily return a maximal matching. Finally, we construct a mechanism that is
universally truthful on path-cycle graphs and whose performance is within 2/3 of optimal.
Again, it is known that no better ratio is possible.
In most of the existing literature, mechanisms are typically evaluated by their overall
performance on a large exchange pool, based on which conclusions and recommendations
are drawn. In our second essay, we consider a dynamic framework to evaluate extensively
used kidney exchange mechanisms. We conduct a simulation-based study of a dynamically
evolving exchange pool during 9 years. Our results suggest that some of the features that
are critical in a mechanism in the static setting have only a minor impact in its longrun
performance when viewed in the dynamic setting. More importantly, features that
are generally underestimated in the static setting turn to be relevant when we look at
dynamically evolving exchange pool. For example, the pairs' arrival rates. In particular we
provide insights into the eect on the waiting times and the probability to receive an oer
of controllable features such as the frequency at which matching are run, the structures
through which pairs could be matched (cycles or chains) as well as inherent features such
as the pairs ABO-PRA characteristics, the availability of altruistic donors, and wether or
not compatible pairs join the exchange etc. We evaluate the odds to receive an oer and
the expected time to receive an oer for each ABO-PRA type of pairs in the model.
Power management in computing devices aims to minimize energy consumption to perform
tasks, meanwhile keeping acceptable performance levels. A widely used power management
strategy for devices, is to transit the devices and/or components to lower power
consumption states during inactivity periods. Transitions between power states consume
energy, thus, depending on such costs, it may be advantageous to stay in high power state
during some inactivity periods. In our third essay we consider the problem of minimizing
the total energy consumed by a 2-power state device, to process jobs that are sent over time
by a constrained adversary. Jobs can be preempted, but deadlines need to be met. In this
problem, an algorithm must decide when to schedule the jobs, as well as a sequence of power
states, and the discrete time thresholds at which these states will be reached. We provide
an online algorithm to minimize the energy consumption when the cost of a transition to
the low power state is small enough. In this case, the problem of minimizing the energy
consumption is equivalent to minimizing the total number of inactivity periods. We also
provide an algorithm to minimize the energy consumption when it may be advantageous
to stay in high power state during some inactivity periods. In both cases we provide upper
bounds on the competitive ratio of our algorithms, and lower bounds on the competitive
ratio of all online algorithms.Operations researchIndustrial Engineering and Operations ResearchThesesSequential Optimization in Changing Environments: Theory and Application to Online Content Recommendation Services
https://academiccommons.columbia.edu/catalog/ac:176086
Gur, Yonatan10.7916/D8639MWFThu, 08 Jun 2017 20:14:02 +0000Recent technological developments allow the online collection of valuable information that can be efficiently used to optimize decisions "on the fly" and at a low cost. These advances have greatly influenced the decision-making process in various areas of operations management, including pricing, inventory, and retail management. In this thesis we study methodological as well as practical aspects arising in online sequential optimization in the presence of such real-time information streams. On the methodological front, we study aspects of sequential optimization in the presence of temporal changes, such as designing decision making policies that adopt to temporal changes in the underlying environment (that drives performance) when only partial information about this changing environment is available, and quantifying the added complexity in sequential decision making problems when temporal changes are introduced. On the applied front, we study practical aspects associated with a class of online services that focus on creating customized recommendations (e.g., Amazon, Netflix). In particular, we focus on online content recommendations, a new class of online services that allows publishers to direct readers from articles they are currently reading to other web-based content they may be interested in, by means of links attached to said article.
In the first part of the thesis we consider a non-stationary variant of a sequential stochastic optimization problem, where the underlying cost functions may change along the horizon. We propose a measure, termed {\it variation budget}, that controls the extent of said change, and study how restrictions on this budget impact achievable performance. As a yardstick to quantify performance in non-stationary settings we propose a regret measure relative to a dynamic oracle benchmark. We identify sharp conditions under which it is possible to achieve long-run-average optimality and more refined performance measures such as rate optimality that fully characterize the complexity of such problems. In doing so, we also establish a strong connection between two rather disparate strands of literature: adversarial online convex optimization; and the more traditional stochastic approximation paradigm (couched in a non-stationary setting). This connection is the key to deriving well performing policies in the latter, by leveraging structure of optimal policies in the former. Finally, tight bounds on the minimax regret allow us to quantify the "price of non-stationarity," which mathematically captures the added complexity embedded in a temporally changing environment versus a stationary one.
In the second part of the thesis we consider another core stochastic optimization problem couched in a multi-armed bandit (MAB) setting. We develop a MAB formulation that allows for a broad range of temporal uncertainties in the rewards, characterize the (regret) complexity of this class of MAB problems by establishing a direct link between the extent of allowable reward "variation" and the minimal achievable worst-case regret, and provide an optimal policy that achieves that performance. Similarly to the first part of the thesis, our analysis draws concrete connections between two strands of literature: the adversarial and the stochastic MAB frameworks.
The third part of the thesis studies applied optimization aspects arising in online content recommendations, that allow web-based publishers to direct readers from articles they are currently reading to other web-based content. We study the content recommendation problem and its unique dynamic features from both theoretical as well as practical perspectives. Using a large data set of browsing history at major media sites, we develop a representation of content along two key dimensions: clickability, the likelihood to click to an article when it is recommended; and engageability, the likelihood to click from an article when it hosts a recommendation. Based on this representation, we propose a class of user path-focused heuristics, whose purpose is to simultaneously ensure a high instantaneous probability of clicking recommended articles, while also optimizing engagement along the future path. We rigorously quantify the performance of these heuristics and validate their impact through a live experiment. The third part of the thesis is based on a collaboration with a leading provider of content recommendations to online publishers.Operations research, Business, MathematicsBusinessThesesStudies in Stochastic Networks: Efficient Monte-Carlo Methods, Modeling and Asymptotic Analysis
https://academiccommons.columbia.edu/catalog/ac:177127
Dong, Jing10.7916/D8X63K4FThu, 08 Jun 2017 20:11:58 +0000This dissertation contains two parts. The first part develops a series of bias reduction techniques for: point processes on stable unbounded regions, steady-state distribution of infinite server queues, steady-state distribution of multi-server loss queues and loss networks and sample path of stochastic differential equations. These techniques can be applied for efficient performance evaluation and optimization of the corresponding stochastic models. We perform detailed running time analysis under heavy traffic of the perfect sampling algorithms for infinite server queues and multi-server loss queues and prove that the algorithms achieve nearly optimal order of complexity. The second part aims to model and analyze the load-dependent slowdown effect in service systems. One important phenomenon we observe in such systems is bi-stability, where the system alternates randomly between two performance regions. We conduct heavy traffic asymptotic analysis of system dynamics and provide operational solutions to avoid the bad performance region.Operations research, Mathematicsjd2736Industrial Engineering and Operations ResearchThesesNetwork Resource Allocation Under Fairness Constraints
https://academiccommons.columbia.edu/catalog/ac:176038
Chandramouli, Shyam Sundar10.7916/D8S46Q3VThu, 08 Jun 2017 20:11:00 +0000This work considers the basic problem of allocating resources among a group of agents in a network, when the agents are equipped with single-peaked preferences over their assignments. This generalizes the classical claims problem, which concerns the division of an estate's liquidation value when the total claim on it exceeds this value. The claims problem also models the problem of rationing a single commodity, or the problem of dividing the cost of a public project among the people it serves, or the problem of apportioning taxes. A key consideration in this classical literature is equity: the good (or the ``bad,'' in the case of apportioning taxes or costs) should be distributed as fairly as possible. The main contribution of this dissertation is a comprehensive treatment of a generalization of this classical rationing problem to a network setting.
Bochet et al. recently introduced a generalization of the classical rationing problem to the network setting. For this problem they designed an allocation mechanism---the egalitarian mechanism---that is Pareto optimal, envy free and strategyproof. In chapter 2, it is shown that the egalitarian mechanism is in fact group strategyproof, implying that no coalition of agents can collectively misreport their information to obtain a (weakly) better allocation for themselves. Further, a complete characterization of the set of all group strategyproof mechanisms is obtained.
The egalitarian mechanism satisfies many attractive properties, but fails consistency, an important property in the literature on rationing problems. It is shown in chapter 3 that no Pareto optimal mechanism can be envy-free and consistent. Chapter 3 is devoted to the edge-fair mechanism that is Pareto optimal, group strategyproof, and consistent. In a related model where the agents are located on the edges of the graph rather than the nodes, the edge-fair rule is shown to be envy-free, group strategyproof, and consistent.
Chapter 4 extends the egalitarian mechanism to the problem of finding an optimal exchange in non-bipartite networks. The results vary depending on whether the commodity being exchanged is divisible or indivisible. For the latter case, it is shown that no efficient mechanism can be strategyproof, and that the egalitarian mechanism is Pareto optimal and envy-free. Chapter 5 generalizes recent work on finding stable and balanced allocations in graphs with unit capacities and unit weights to more general networks. The existence of a stable and balanced allocation is established by a transformation to an equivalent unit capacity network.Operations researchIndustrial Engineering and Operations ResearchThesesEssays in Financial Engineering
https://academiccommons.columbia.edu/catalog/ac:177072
Ahn, Andrew10.7916/D80K26R0Thu, 08 Jun 2017 20:08:08 +0000This thesis consists of three essays in financial engineering. In particular we study problems in option pricing, stochastic control and risk management.
In the first essay, we develop an accurate and efficient pricing approach for options on leveraged ETFs (LETFs). Our approach allows us to price these options quickly and in a manner that is consistent with the underlying ETF price dynamics. The numerical results also demonstrate that LETF option prices have model-dependency particularly in high-volatility environments.
In the second essay, we extend a linear programming (LP) technique for approximately solving high-dimensional control problems in a diffusion setting. The original LP technique applies to finite horizon problems with an exponentially-distributed horizon, T. We extend the approach to fixed horizon problems. We then apply these techniques to dynamic portfolio optimization problems and evaluate their performance using convex duality methods. The numerical results suggest that the LP approach is a very promising one for tackling high-dimensional control problems.
In the final essay, we propose a factor model-based approach for performing scenario analysis in a risk management context. We argue that our approach addresses some important drawbacks to a standard scenario analysis and, in a preliminary numerical investigation with option portfolios, we show that it produces superior results as well.Operations researchaja2133Industrial Engineering and Operations ResearchThesesPricing, Trading and Clearing of Defaultable Claims Subject to Counterparty Risk
https://academiccommons.columbia.edu/catalog/ac:169814
Kim, Jinbeom10.7916/D8319SWWThu, 08 Jun 2017 16:12:44 +0000The recent financial crisis and subsequent regulatory changes on over-the-counter (OTC) markets have given rise to the new valuation and trading frameworks for defaultable claims to investors and dealer banks. More OTC market participants have adopted the new market conventions that incorporate counterparty risk into the valuation of OTC derivatives. In addition, the use of collateral has become common for most bilateral trades to reduce counterparty default risk. On the other hand, to increase transparency and market stability, the U.S and European regulators have required mandatory clearing of defaultable derivatives through central counterparties. This dissertation tackles these changes and analyze their impacts on the pricing, trading and clearing of defaultable claims. In the first part of the thesis, we study a valuation framework for financial contracts subject to reference and counterparty default risks with collateralization requirement. We propose a fixed point approach to analyze the mark-to-market contract value with counterparty risk provision, and show that it is a unique bounded and continuous fixed point via contraction mapping. This leads us to develop an accurate iterative numerical scheme for valuation. Specifically, we solve a sequence of linear inhomogeneous partial differential equations, whose solutions converge to the fixed point price function. We apply our methodology to compute the bid and ask prices for both defaultable equity and fixed-income derivatives, and illustrate the non-trivial effects of counterparty risk, collateralization ratio and liquidation convention on the bid-ask prices. In the second part, we study the problem of pricing and trading of defaultable claims among investors with heterogeneous risk preferences and market views. Based on the utility-indifference pricing methodology, we construct the bid-ask spreads for risk-averse buyers and sellers, and show that the spreads widen as risk aversion or trading volume increases. Moreover, we analyze the buyer's optimal static trading position under various market settings, including (i) when the market pricing rule is linear, and (ii) when the counterparty -- single or multiple sellers -- may have different nonlinear pricing rules generated by risk aversion and belief heterogeneity. For defaultable bonds and credit default swaps, we provide explicit formulas for the optimal trading positions, and examine the combined effect of heterogeneous risk aversions and beliefs. In particular, we find that belief heterogeneity, rather than the difference in risk aversion, is crucial to trigger a trade. Finally, we study the impact of central clearing on the credit default swap (CDS) market. Central clearing of CDS through a central counterparty (CCP) has been proposed as a tool for mitigating systemic risk and counterpart risk in the CDS market. The design of CCPs involves the implementation of margin requirements and a default fund, for which various designs have been proposed. We propose a mathematical model to quantify the impact of the design of the CCP on the incentive for clearing and analyze the market equilibrium. We determine the minimum number of clearing participants required so that they have an incentive to clear part of their exposures. Furthermore, we analyze the equilibrium CDS positions and their dependence on the initial margin, risk aversion, and counterparty risk in the inter-dealer market. Our numerical results show that minimizing the initial margin maximizes the total clearing positions as well as the CCP's revenue.Operations research, Financejk3071Industrial Engineering and Operations ResearchThesesPerfect Simulation, Sample-path Large Deviations, and Multiscale Modeling for Some Fundamental Queueing Systems
https://academiccommons.columbia.edu/catalog/ac:181094
Chen, Xinyun10.7916/D8WH2MZ1Thu, 08 Jun 2017 16:12:43 +0000As a primary branch of Operations Research, Queueing Theory models and analyzes engineering systems with random fluctuations. With the development of internet and computation techniques, the engineering systems today are much bigger in scale and more complicated in structure than 20 years ago, which raises numerous new problems to researchers in the field of queueing theory. The aim of this thesis is to explore new methods and tools, from both algorithmic and analytical perspectives, that are useful to solve such problems.
In Chapter 1 and 2, we introduce some techniques of asymptotic analysis that are relatively new to queueing applications in order to give more accurate probabilistic characterization of queueing models with large scale and complicated structure. In particular, Chapter 1 gives the first functional large deviation result for infinite-server system with general inter-arrival and service times. The functional approach we use enables a nice description of the whole system over the entire time horizon of interest, which is important in real problems. In Chapter 2, we construct a queueing model for the so-called limit order book that is used in main financial markets worldwide. We use an asymptotic approach called multi-scale modeling to disentangle the complicated dependence among the elements in the trading system and to reduce the model dimensionality. The asymptotic regime we use is inspired by empirical observations and the resulting limit process explains and reproduces stylized features of real market data. Chapter 2 also provides a nice example of novel applications of queueing models in systems, such as the electronic trading system, that are traditionally outside the scope of queueing theory.
Chapter 3 and 4 focus on stochastic simulation methods for performance evaluation of queueing models where analytic approaches fail.
In Chapter 3, we develop a perfect sampling algorithm to generate exact samples from the stationary distribution of stochastic fluid networks in polynomial time. Our approach can be used for time-varying networks with general inter-arrival and service times, whose stationary distributions have no analytic expression. In Chapter 4, we focus on the stochastic systems with continuous random fluctuations, for instance, the workload arrives to the system in continuous flow like a Levy process. We develop a general framework of simulation algorithms featuring a deterministic error bound and an almost square root convergence rate. As an application, we apply this framework to estimate the stationary distributions of reflected Brownian motions and the performance of our algorithm is better than existing prevalent numeric methods.Operations researchxc2177Industrial Engineering and Operations ResearchThesesApproximate dynamic programming for large scale systems
https://academiccommons.columbia.edu/catalog/ac:169790
Desai, Vijay V.10.7916/D82V2PH1Thu, 08 Jun 2017 16:12:40 +0000Sequential decision making under uncertainty is at the heart of a wide variety of practical problems. These problems can be cast as dynamic programs and the optimal value function can be computed by solving Bellman's equation. However, this approach is limited in its applicability. As the number of state variables increases, the state space size grows exponentially, a phenomenon known as the curse of dimensionality, rendering the standard dynamic programming approach impractical. An effective way of addressing curse of dimensionality is through parameterized value function approximation. Such an approximation is determined by relatively small number of parameters and serves as an estimate of the optimal value function. But in order for this approach to be effective, we need Approximate Dynamic Programming (ADP) algorithms that can deliver `good' approximation to the optimal value function and such an approximation can then be used to derive policies for effective decision-making. From a practical standpoint, in order to assess the effectiveness of such an approximation, there is also a need for methods that give a sense for the suboptimality of a policy. This thesis is an attempt to address both these issues. First, we introduce a new ADP algorithm based on linear programming, to compute value function approximations. LP approaches to approximate DP have typically relied on a natural `projection' of a well studied linear program for exact dynamic programming. Such programs restrict attention to approximations that are lower bounds to the optimal cost-to-go function. Our program -- the `smoothed approximate linear program' -- is distinct from such approaches and relaxes the restriction to lower bounding approximations in an appropriate fashion while remaining computationally tractable. The resulting program enjoys strong approximation guarantees and is shown to perform well in numerical experiments with the game of Tetris and queueing network control problem. Next, we consider optimal stopping problems with applications to pricing of high-dimensional American options. We introduce the pathwise optimization (PO) method: a new convex optimization procedure to produce upper and lower bounds on the optimal value (the `price') of high-dimensional optimal stopping problems. The PO method builds on a dual characterization of optimal stopping problems as optimization problems over the space of martingales, which we dub the martingale duality approach. We demonstrate via numerical experiments that the PO method produces upper bounds and lower bounds (via suboptimal exercise policies) of a quality comparable with state-of-the-art approaches. Further, we develop an approximation theory relevant to martingale duality approaches in general and the PO method in particular. Finally, we consider a broad class of MDPs and introduce a new tractable method for computing bounds by consider information relaxation and introducing penalty. The method delivers tight bounds by identifying the best penalty function among a parameterized class of penalty functions. We implement our method on a high-dimensional financial application, namely, optimal execution and demonstrate the practical value of the method vis-a-vis competing methods available in the literature. In addition, we provide theory to show that bounds generated by our method are provably tighter than some of the other available approaches.Operations research, Mathematicsvvd2101Industrial Engineering and Operations ResearchThesesTwo Papers of Financial Engineering Relating to the Risk of the 2007--2008 Financial Crisis
https://academiccommons.columbia.edu/catalog/ac:167143
Zhong, Haowen10.7916/D8CC0XMGThu, 08 Jun 2017 16:12:33 +0000This dissertation studies two financial engineering and econometrics problems relating to two facets of the 2007-2008 financial crisis. In the first part, we construct the Spatial Capital Asset Pricing Model and the Spatial Arbitrage Pricing Theory to characterize the risk premiums of futures contracts on real estate assets. We also provide rigorous econometric analysis of the new models. Empirical study shows there exists significant spatial interaction among the S&P/Case-Shiller Home Price Index futures returns. In the second part, we perform empirical studies on the jump risk in the equity market. We propose a simple affine jump-diffusion model for equity returns, which seems to outperform existing ones (including models with Levy jumps) during the financial crisis and is at least as good during normal times, if model complexity is taken into account. In comparing the models, we made two empirical findings: (i) jump intensity seems to increase significantly during the financial crisis, while on average there appears to be little change of jump sizes; (ii) finite number of large jumps in returns for any finite time horizon seem to fit the data well both before and after the crisis.Operations research, Statisticshz2193Industrial Engineering and Operations ResearchThesesFrom Continuous to Discrete: Studies on Continuity Corrections and Monte Carlo Simulation with Applications to Barrier Options and American Options
https://academiccommons.columbia.edu/catalog/ac:171186
Cao, Menghui10.7916/D8PG1PS1Thu, 08 Jun 2017 16:12:23 +0000This dissertation 1) shows continuity corrections for first passage probabilities of Brownian bridge and barrier joint probabilities, which are applied to the pricing of two-dimensional barrier and partial barrier options, and 2) introduces new variance reduction techniques and computational improvements to Monte Carlo methods for pricing American options.
The joint distribution of Brownian motion and its first passage time has found applications in many areas, including sequential analysis, pricing of barrier options, and credit risk modeling. There are, however, no simple closed-form solutions for these joint probabilities in a discrete-time setting. Chapter 2 shows that, discrete two-dimensional barrier and partial barrier joint probabilities can be approximated by their continuous-time probabilities with remarkable accuracy after shifting the barrier away from the underlying by a factor. We achieve this through a uniform continuity correction theorem on the first passage probabilities for Brownian bridge, extending relevant results in Siegmund (1985a). The continuity corrections are applied to the pricing of two-dimensional barrier and partial barrier options, extending the results in Broadie, Glasserman & Kou (1997) on one-dimensional barrier options. One interesting aspect is that for type B partial barrier options, the barrier correction cannot be applied throughout one pricing formula, but only to some barrier values and leaving the other unchanged, the direction of correction may also vary within one formula.
In Chapter 3 we introduce new variance reduction techniques and computational improvements to Monte Carlo methods for pricing American-style options. For simulation algorithms that compute lower bounds of American option values, we apply martingale control variates and introduce the local policy enhancement, which adopts a local simulation to improve the exercise policy. For duality-based upper bound methods, specifically the primal-dual simulation algorithm (Andersen and Broadie 2004), we have developed two improvements. One is sub-optimality checking, which saves unnecessary computation when it is sub-optimal to exercise the option along the sample path; the second is boundary distance grouping, which reduces computational time by skipping computation on selected sample paths based on the distance to the exercise boundary. Numerical results are given for single asset Bermudan options, moving window Asian options and Bermudan max options. In some examples the computational time is reduced by a factor of several hundred, while the confidence interval of the true option value is considerably tighter than before the improvements.Operations research, FinanceIndustrial Engineering and Operations ResearchThesesDynamic Markets with Many Agents: Applications in Social Learning and Competition
https://academiccommons.columbia.edu/catalog/ac:174798
Ifrach, Bar10.7916/D8NK3C48Thu, 08 Jun 2017 16:06:11 +0000This thesis considers two applications in dynamics economic models with many agents. The dynamics of the economic systems under consideration are intractable since they depend on the (stochastic) outcomes of the agents' actions. However, as the number of agents grows large, approximations to the aggregate behavior of agents come to light. I use this observation to characterize market dynamics and subsequently to study these applications.
Chapter 2 studies the problem of devising a pricing strategy to maximize the revenues extracted from a stream of consumers with heterogenous preferences. Consumers, however, do not know the quality of the product or service and engage in a social learning process to learn it. Using a mean-field approximation the transient of this social learning process is uncovered and the pricing problem is analyzed.
Chapter 3 adds to the previous chapter in analyzing features of this social learning process with finitely many agents. In addition, the chapter generalizes the information structure to include cases where consumers take into account the order in which reviews were submitted.
Chapter 4 considers a model of dynamic oligopoly competition in the spirit of models that are widespread in industrial organization. The computation of equilibrium strategies of such models suffers from the curse of dimensionality when the number of agents (firms) is large. For a market structure with few dominant firms and many fringe firms, I study an alternative equilibrium concept in which fringe firms are represented succinctly with a low dimensional set of statistics. The chapter explores how this new equilibrium concept expands the class of dynamic oligopoly models that can be studied computationally in empirical work.Operations research, Economicsbi2118BusinessThesesData-driven System Design in Service Operations
https://academiccommons.columbia.edu/catalog/ac:163306
Lu, Yina10.7916/D8WM1MSFThu, 08 Jun 2017 16:06:11 +0000The service industry has become an increasingly important component in the world's economy. Simultaneously, the data collected from service systems has grown rapidly in both size and complexity due to the rapid spread of information technology, providing new opportunities and challenges for operations management researchers. This dissertation aims to explore methodologies to extract information from data and provide powerful insights to guide the design of service delivery systems. To do this, we analyze three applications in the retail, healthcare, and IT service industries. In the first application, we conduct an empirical study to analyze how waiting in queue in the context of a retail store affects customers' purchasing behavior. The methodology combines a novel dataset collected via video recognition technology with traditional point-of-sales data. We find that waiting in queue has a nonlinear impact on purchase incidence and that customers appear to focus mostly on the length of the queue, without adjusting enough for the speed at which the line moves. We also find that customers' sensitivity to waiting is heterogeneous and negatively correlated with price sensitivity. These findings have important implications for queueing system design and pricing management under congestion. The second application focuses on disaster planning in healthcare. According to a U.S. government mandate, in a catastrophic event, the New York City metropolitan areas need to be capable of caring for 400 burn-injured patients during a catastrophe, which far exceeds the current burn bed capacity. We develop a new system for prioritizing patients for transfer to burn beds as they become available and demonstrate its superiority over several other triage methods. Based on data from previous burn catastrophes, we study the feasibility of being able to admit the required number of patients to burn beds within the critical three-to-five-day time frame. We find that this is unlikely and that the ability to do so is highly dependent on the type of event and the demographics of the patient population. This work has implications for how disaster plans in other metropolitan areas should be developed. In the third application, we study workers' productivity in a global IT service delivery system, where service requests from possibly globally distributed customers are managed centrally and served by agents. Based on a novel dataset which tracks the detailed time intervals an agent spends on all business related activities, we develop a methodology to study the variation of productivity over time motivated by econometric tools from survival analysis. This approach can be used to identify different mechanisms by which workload affects productivity. The findings provide important insights for the design of the workload allocation policies which account for agents' workload management behavior.Operations researchyl2494BusinessThesesDesign and Evaluation of Procurement Combinatorial Auctions
https://academiccommons.columbia.edu/catalog/ac:173476
Kim, Sang Won10.7916/D8DF6P8CThu, 08 Jun 2017 16:06:09 +0000The main advantage of a procurement combinatorial auction (CA) is that it allows suppliers to express cost synergies through package bids. However, bidders can also strategically take advantage of this flexibility, by discounting package bids and "inflating" bid prices for single-items, even in the absence of cost synergies; the latter behavior can hurt the performance of the auction. It is an empirical question whether allowing package bids and running a CA improves performance in a given setting.Analyzing the actual performance of a CA requires evaluating cost efficiency and the margins of the winning bidders, which is typically private and sensitive information of the bidders. Thus motivated, in Chapter 2 of this dissertation, we develop a structural estimation approach for large-scale first-price CAs to estimate the firms' cost structure using the bid data. To overcome the computational difficulties arising from the large number of bids observed in large-scale CAs, we propose a novel simplified model of bidders' behavior based on pricing package characteristics. Overall, this work develops the first practical tool to empirically evaluate the performance of large-scale first-price CAs commonly used in procurement settings.In Chapter 3, we apply our method to the Chilean school meals auction, in which the government procures half a billion dollars' worth of meal services every year and bidders submit thousands of package bids. Our estimates suggest that bidders' cost synergies are economically significant in this application (~5%), and the current CA mechanism achieves high allocative efficiency (~98%) and reasonable margins for the bidders (~5%). We believe this is the first work in the literature that empirically shows that a CA performs well in a real-world application.We also conduct a counterfactual analysis to study the performance of the Vickrey-Clarke-Groves (VCG) mechanism in our empirical application. While it is well known in the literature that the VCG mechanism achieves allocative efficiency, its application in practice is at best rare due to several potential weaknesses such as prohibitively high procurement costs. Interestingly, contrary to the recent theoretical work, the results show that the VCG mechanism achieves reasonable procurement costs in our application. Motivated from this observation, Chapter 4 addresses such apparent paradox between the theory and our empirical application. Focusing on the high procurement cost issue, we study the impact of competition on the revenue performance of the VCG mechanism using an asymptotic analysis. We believe the findings in this chapter add useful insights for the practical usage of the VCG mechanism.Business, Economics, Operations researchskim14BusinessThesesInfrastructure Scaling and Pricing
https://academiccommons.columbia.edu/catalog/ac:171000
Gocmen, Fikret Caner10.7916/D8SQ8XFFThu, 08 Jun 2017 16:06:09 +0000Infrastructure systems play a crucial role in our daily lives. They include, but are not limited to, the highways we take while we commute to work, the stadiums we go to watch games, and the power plants that provide the electricity we consume in our homes. In this thesis we study infrastructure systems from several different perspectives with a focus on pricing and scalability. The pricing aspect of our research focuses on two industries: toll roads and sports events. Afterwards, we analyze the potential impact of small modular infrastructure on a wide variety of industries. We start by analyzing the problem of determining the tolls that maximize revenue for a managed lane operator -- that is, an operator who can charge a toll for the use of some lanes on a highway while a number of parallel lanes remain free to use. Managing toll lanes for profit is becoming increasingly common as private contractors agree to build additional lane capacity in return for the opportunity to retain toll revenue. We start by modeling the lanes as queues and show that the dynamic revenue-maximizing toll is always greater than or equal to the myopic toll that maximizes expected revenue from each arriving vehicle. Numerical examples show that a dynamic revenue-maximizing toll scheme can generate significantly more expected revenue than either a myopic or a static toll scheme. An important implication is that the revenue-maximizing fee does not only depend on the current state, but also on anticipated future arrivals. We discuss the managerial implications and present several numerical examples. Next, we relax the queueing assumption and model traffic propagation on a highway realistically by using simulation. We devise a framework that can be used to obtain revenue maximizing tolls in such a context. We calibrate our framework by using data from the SR-91 Highway in Orange County, CA and explore different tolling schemes. Our numerical experiments suggest that simple dynamic tolling mechanisms can lead to substantial revenue improvements over myopic and time-of-use tolling policies. In the third part, we analyze the revenue management of consumer options for tournaments. Sporting event managers typically only offer advance tickets which guarantee a seat at a future sporting event in return for an upfront payment. Some event managers and ticket resellers have started to offer call options under which a customer can pay a small amount now for the guaranteed option to attend a future sporting event by paying an additional amount later. We consider the case of tournament options where the event manager sells team-specific options for a tournament final, such as the Super Bowl, before the finalists are determined. These options guarantee a final game ticket to the bearer if his team advances to the finals. We develop an approach by which an event manager can determine the revenue maximizing prices and amounts of advance tickets and options to sell for a tournament final. Afterwards, for a specific tournament structure we show that offering options is guaranteed to increase expected revenue for the event. We also establish bounds for the revenue improvement and show that introducing options can increase social welfare. We conclude by presenting a numerical application of our approach. Finally, we argue that advances made in automation, communication and manufacturing portend a dramatic reversal of the ``bigger is better'' approach to cost reductions prevalent in many basic infrastructure industries, e.g. transportation, electric power generation and raw material processing. We show that the traditional reductions in capital costs achieved by scaling up in size are generally matched by learning effects in the mass-production process when scaling up in numbers instead. In addition, using the U.S. electricity generation sector as a case study, we argue that the primary operating cost advantage of large unit scale is reduced labor, which can be eliminated by employing low-cost automation technologies. Finally, we argue that locational, operational and financial flexibilities that accompany smaller unit scale can reduce investment and operating costs even further. All these factors combined argue that with current technology, economies of numbers may well dominate economies of unit scale.Business, Operations research, MathematicsBusinessThesesOptimization Algorithms for Structured Machine Learning and Image Processing Problems
https://academiccommons.columbia.edu/catalog/ac:158764
Qin, Zhiwei10.7916/D8JH3TDMThu, 08 Jun 2017 13:55:12 +0000Optimization algorithms are often the solution engine for machine learning and image processing techniques, but they can also become the bottleneck in applying these techniques if they are unable to cope with the size of the data. With the rapid advancement of modern technology, data of unprecedented size has become more and more available, and there is an increasing demand to process and interpret the data. Traditional optimization methods, such as the interior-point method, can solve a wide array of problems arising from the machine learning domain, but it is also this generality that often prevents them from dealing with large data efficiently. Hence, specialized algorithms that can readily take advantage of the problem structure are highly desirable and of immediate practical interest. This thesis focuses on developing efficient optimization algorithms for machine learning and image processing problems of diverse types, including supervised learning (e.g., the group lasso), unsupervised learning (e.g., robust tensor decompositions), and total-variation image denoising. These algorithms are of wide interest to the optimization, machine learning, and image processing communities. Specifically, (i) we present two algorithms to solve the Group Lasso problem. First, we propose a general version of the Block Coordinate Descent (BCD) algorithm for the Group Lasso that employs an efficient approach for optimizing each subproblem exactly. We show that it exhibits excellent performance when the groups are of moderate size. For groups of large size, we propose an extension of the proximal gradient algorithm based on variable step-lengths that can be viewed as a simplified version of BCD. By combining the two approaches we obtain an implementation that is very competitive and often outperforms other state-of-the-art approaches for this problem. We show how these methods fit into the globally convergent general block coordinate gradient descent framework in (Tseng and Yun, 2009). We also show that the proposed approach is more efficient in practice than the one implemented in (Tseng and Yun, 2009). In addition, we apply our algorithms to the Multiple Measurement Vector (MMV) recovery problem, which can be viewed as a special case of the Group Lasso problem, and compare their performance to other methods in this particular instance; (ii) we further investigate sparse linear models with two commonly adopted general sparsity-inducing regularization terms, the overlapping Group Lasso penalty l1/l2-norm and the l1/l_infty-norm. We propose a unified framework based on the augmented Lagrangian method, under which problems with both types of regularization and their variants can be efficiently solved. As one of the core building-blocks of this framework, we develop new algorithms using a partial-linearization/splitting technique and prove that the accelerated versions of these algorithms require $O(1 sqrt(epsilon) ) iterations to obtain an epsilon-optimal solution. We compare the performance of these algorithms against that of the alternating direction augmented Lagrangian and FISTA methods on a collection of data sets and apply them to two real-world problems to compare the relative merits of the two norms; (iii) we study the problem of robust low-rank tensor recovery in a convex optimization framework, drawing upon recent advances in robust Principal Component Analysis and tensor completion. We propose tailored optimization algorithms with global convergence guarantees for solving both the constrained and the Lagrangian formulations of the problem. These algorithms are based on the highly efficient alternating direction augmented Lagrangian and accelerated proximal gradient methods. We also propose a nonconvex model that can often improve the recovery results from the convex models. We investigate the empirical recoverability properties of the convex and nonconvex formulations and compare the computational performance of the algorithms on simulated data. We demonstrate through a number of real applications the practical effectiveness of this convex optimization framework for robust low-rank tensor recovery; (iv) we consider the image denoising problem using total variation regularization. This problem is computationally challenging to solve due to the non-differentiability and non-linearity of the regularization term. We propose a new alternating direction augmented Lagrangian method, involving subproblems that can be solved efficiently and exactly. The global convergence of the new algorithm is established for the anisotropic total variation model. We compare our method with the split Bregman method and demonstrate the superiority of our method in computational performance on a set of standard test images.Operations research, Computer science, Statisticszq2107Industrial Engineering and Operations ResearchThesesFinancial Portfolio Risk Management: Model Risk, Robustness and Rebalancing Error
https://academiccommons.columbia.edu/catalog/ac:161415
Xu, Xingbo10.7916/D8SX6MF1Thu, 08 Jun 2017 13:54:56 +0000Risk management has always been in key component of portfolio management. While more and more complicated models are proposed and implemented as research advances, they all inevitably rely on imperfect assumptions and estimates. This dissertation aims to investigate the gap between complicated theoretical modelling and practice. We mainly focus on two directions: model risk and reblancing error. In the first part of the thesis, we develop a framework for quantifying the impact of model error and for measuring and minimizing risk in a way that is robust to model error. This robust approach starts from a baseline model and finds the worst-case error in risk measurement that would be incurred through a deviation from the baseline model, given a precise constraint on the plausibility of the deviation. Using relative entropy to constrain model distance leads to an explicit characterization of worst-case model errors; this characterization lends itself to Monte Carlo simulation, allowing straightforward calculation of bounds on model error with very little computational effort beyond that required to evaluate performance under the baseline nominal model. This approach goes well beyond the effect of errors in parameter estimates to consider errors in the underlying stochastic assumptions of the model and to characterize the greatest vulnerabilities to error in a model. We apply this approach to problems of portfolio risk measurement, credit risk, delta hedging, and counterparty risk measured through credit valuation adjustment. In the second part, we apply this robust approach to a dynamic portfolio control problem. The sources of model error include the evolution of market factors and the influence of these factors on asset returns. We analyze both finite- and infinite-horizon problems in a model in which returns are driven by factors that evolve stochastically. The model incorporates transaction costs and leads to simple and tractable optimal robust controls for multiple assets. We illustrate the performance of the controls on historical data. Robustness does improve performance in out-of-sample tests in which the model is estimated on a rolling window of data and then applied over a subsequent time period. By acknowledging uncertainty in the estimated model, the robust rules lead to less aggressive trading and are less sensitive to sharp moves in underlying prices. In the last part, we analyze the error between a discretely rebalanced portfolio and its continuously rebalanced counterpart in the presence of jumps or mean-reversion in the underlying asset dynamics. With discrete rebalancing, the portfolio's composition is restored to a set of fixed target weights at discrete intervals; with continuous rebalancing, the target weights are maintained at all times. We examine the difference between the two portfolios as the number of discrete rebalancing dates increases. We derive the limiting variance of the relative error between the two portfolios for both the mean-reverting and jump-diffusion cases. For both cases, we derive ``volatility adjustments'' to improve the approximation of the discretely rebalanced portfolio by the continuously rebalanced portfolio, based on on the limiting covariance between the relative rebalancing error and the level of the continuously rebalanced portfolio. These results are based on strong approximation results for jump-diffusion processes.Operations research, Finance, Mathematicsxx2126Industrial Engineering and Operations ResearchThesesStochastic Models of Limit Order Markets
https://academiccommons.columbia.edu/catalog/ac:161685
Kukanov, Arseniy10.7916/D8C253MBThu, 08 Jun 2017 13:53:49 +0000During the last two decades most stock and derivatives exchanges in the world transitioned to electronic trading in limit order books, creating a need for a new set of quantitative models to describe these order-driven markets. This dissertation offers a collection of models that provide insight into the structure of modern financial markets, and can help to optimize trading decisions in practical applications. In the first part of the thesis we study the dynamics of prices, order flows and liquidity in limit order markets over short timescales. We propose a stylized order book model that predicts a particularly simple linear relation between price changes and order flow imbalance, defined as a difference between net changes in supply and demand. The slope in this linear relation, called a price impact coefficient, is inversely proportional in our model to market depth - a measure of liquidity. Our empirical results confirm both of these predictions. The linear relation between order flow imbalance and price changes holds for time intervals between 50 milliseconds and 5 minutes. The inverse relation between the price impact coefficient and market depth holds on longer timescales. These findings shed a new light on intraday variations in market volatility. According to our model volatility fluctuates due to changes in market depth or in order flow variance. Previous studies also found a positive correlation between volatility and trading volume, but in order-driven markets prices are determined by the limit order book activity, so the association between trading volume and volatility is unclear. We show how a spurious correlation between these variables can indeed emerge in our linear model due to time aggregation of high-frequency data. Finally, we observe short-term positive autocorrelation in order flow imbalance and discuss an application of this variable as a measure of adverse selection in limit order executions. Our results suggest that monitoring recent order flow can improve the quality of order executions in practice. In the second part of the thesis we study the problem of optimal order placement in a fragmented limit order market. To execute a trade, market participants can submit limit orders or market orders across various exchanges where a stock is traded. In practice these decisions are influenced by sizes of order queues and by statistical properties of order flows in each limit order book, and also by rebates that exchanges pay for limit order submissions. We present a realistic model of limit order executions and formalize the search for an optimal order placement policy as a convex optimization problem. Based on this formulation we study how various factors determine investor's order placement decisions. In a case when a single exchange is used for order execution, we derive an explicit formula for the optimal limit and market order quantities. Our solution shows that the optimal split between market and limit orders largely depends on one's tolerance to execution risk. Market orders help to alleviate this risk because they execute with certainty. Correspondingly, we find that an optimal order allocation shifts to these more expensive orders when the execution risk is of primary concern, for example when the intended trade quantity is large or when it is costly to catch up on the quantity after limit order execution fails. We also characterize the optimal solution in the general case of simultaneous order placement on multiple exchanges, and show that it sets execution shortfall probabilities to specific threshold values computed with model parameters. Finally, we propose a non-parametric stochastic algorithm that computes an optimal solution by resampling historical data and does not require specifying order flow distributions. A numerical implementation of this algorithm is used to study the sensitivity of an optimal solution to changes in model parameters. Our numerical results show that order placement optimization can bring a substantial reduction in trading costs, especially for small orders and in cases when order flows are relatively uncorrelated across trading venues. The order placement optimization framework developed in this thesis can also be used to quantify the costs and benefits of financial market fragmentation from the point of view of an individual investor. For instance, we find that a positive correlation between order flows, which is empirically observed in a fragmented U.S. equity market, increases the costs of trading. As the correlation increases it may become more expensive to trade in a fragmented market than it is in a consolidated market. In the third part of the thesis we analyze the dynamics of limit order queues at the best bid or ask of an exchange. These queues consist of orders submitted by a variety of market participants, yet existing order book models commonly assume that all orders have similar dynamics. In practice, some orders are submitted by trade execution algorithms in an attempt to buy or sell a certain quantity of assets under time constraints, and these orders are canceled if their realized waiting time exceeds a patience threshold. In contrast, high-frequency traders submit and cancel orders depending on the order book state and their orders are not driven by patience. The interaction between these two order types within a single FIFO queue leads bursts of order cancelations for small queues and anomalously long waiting times in large queues. We analyze a fluid model that describes the evolution of large order queues in liquid markets, taking into account the heterogeneity between order submission and cancelation strategies of different traders. Our results show that after a finite initial time interval, the queue reaches a specific structure where all orders from high-frequency traders stay in the queue until execution but most orders from execution algorithms exceed their patience thresholds and are canceled. This "order crowding" effect has been previously noted by participants in highly liquid stock and futures markets and was attributed to a large participation of high-frequency traders. In our model, their presence creates an additional workload, which increases queue waiting times for new orders. Our analysis of the fluid model leads to waiting time estimates that take into account the distribution of order types in a queue. These estimates are tested against a large dataset of realized limit order waiting times collected by a U.S. equity brokerage firm. The queue composition at a moment of order submission noticeably affects its waiting time and we find that assuming a single order type for all orders in the queue leads to unrealistic results. Estimates that assume instead a mix of heterogeneous orders in the queue are closer to empirical data. Our model for a limit order queue with heterogeneous order types also appears to be interesting from a methodological point of view. It introduces a new type of behavior in a queueing system where one class of jobs has state-dependent dynamics, while others are driven by patience. Although this model is motivated by the analysis of limit order books, it may find applications in studying other service systems with state-dependent abandonments.Operations research, Finance, Statisticsak2870Industrial Engineering and Operations ResearchThesesModels for managing surge capacity in the face of an influenza epidemic
https://academiccommons.columbia.edu/catalog/ac:157364
Zenteno, Ana10.7916/D8G1672SThu, 08 Jun 2017 13:53:11 +0000Influenza pandemics pose an imminent risk to society. Yearly outbreaks already represent heavy social and economic burdens. A pandemic could severely affect infrastructure and commerce through high absenteeism, supply chain disruptions, and other effects over an extended and uncertain period of time. Governmental institutions such as the Center for Disease Prevention and Control (CDC) and the U.S. Department of Health and Human Services (HHS) have issued guidelines on how to prepare for a potential pandemic, however much work still needs to be done in order to meet them. From a planner's perspective, the complexity of outlining plans to manage future resources during an epidemic stems from the uncertainty of how severe the epidemic will be. Uncertainty in parameters such as the contagion rate (how fast the disease spreads) makes the course and severity of the epidemic unforeseeable, exposing any planning strategy to a potentially wasteful allocation of resources. Our approach involves the use of additional resources in response to a robust model of the evolution of the epidemic as to hedge against the uncertainty in its evolution and intensity. Under existing plans, large cities would make use of networks of volunteers, students, and recent retirees, or borrow staff from neighboring communities. Taking into account that such additional resources are likely to be significantly constrained (e.g. in quantity and duration), we seek to produce robust emergency staff commitment levels that work well under different trajectories and degrees of severity of the pandemic. Our methodology combines Robust Optimization techniques with Epidemiology (SEIR models) and system performance modeling. We describe cutting-plane algorithms analogous to generalized Benders' decomposition that prove fast and numerically accurate. Our results yield insights on the structure of optimal robust strategies and on practical rules-of-thumb that can be deployed during the epidemic. To assess the efficacy of our solutions, we study their performance under different scenarios and compare them against other seemingly good strategies through numerical experiments. This work would be particularly valuable for institutions that provide public services, whose operations continuity is critical for a community, especially in view of an event of this caliber. As far as we know, this is the first time this problem is addressed in a rigorous way; particularly we are not aware of any other robust optimization applications in epidemiology.Operations research, Public healthacz2103Industrial Engineering and Operations ResearchThesesPrice competition and the impact of service attributes: Structural estimation and analytical characterizations of equilibrium behavior
https://academiccommons.columbia.edu/catalog/ac:153522
Pierson, Margaret Parker10.7916/D89029WZWed, 07 Jun 2017 17:00:33 +0000This dissertation addresses a number of outstanding, fundamental questions in operations management and industrial organization literature. Operations management literature has a long history of studying the competitive impact of operational, firm-level strategic decisions within oligopoly markets. The first essay reports on an empirical study of an important industry, the drive-thru fast-food industry. We estimate a competition model, derived from an underlying Mixed MultiNomial Logit (MNML) consumer choice model, using detailed empirical data. The main goal is to measure to what extent waiting time performance, along with price levels, brand attributes, geographical and demographic factors, impacts competing firms' market shares. The primary goal of our second essay is to characterize the equilibrium behavior of price competition models with Mixed Multinomial Logit (MMNL) demand functions under affine cost structures. In spite of the huge popularity of MMNL models in both the theoretical and empirical literature, it is not known, in general, whether a Nash equilibrium (in pure strategies) of prices exists, and whether the equilibria can be uniquely characterized as the solutions to the system of First Order Condition (FOC) equations. The third essay, which is the most general in its context, we establish that in the absence of cost efficiencies resulting from a merger, aggregate profits of the merging firms increase as do equilibrium prices for general price competition models with general nonlinear demand and cost functions as long as the models are supermodular, with two additional structural conditions: (i) each firm's profit function is strictly quasi-concave in its own price(s), and (ii) markets are competitive, i.e., in the pre-merger industry, each firm's profits increase when any of his competitors increases his price, unilaterally. Even the equilibrium profits of the remaining firms in the industry increase, while the consumer ends up holding the bag, i.e., consumer welfare declines. As demonstrated by this essay, the answers to these sorts of strategy questions have implications not only for the firms and customers but also the policy makers policing these markets.Operations research, BusinessBusinessThesesContingent Capital: Valuation and Risk Implications Under Alternative Conversion Mechanisms
https://academiccommons.columbia.edu/catalog/ac:152933
Nouri, Behzad10.7916/D80P164KWed, 07 Jun 2017 17:00:14 +0000Several proposals for enhancing the stability of the financial system include requirements that banks hold some form of contingent capital, meaning equity that becomes available to a bank in the event of a crisis or financial distress. Specific proposals vary in their choice of conversion trigger and conversion mechanism, and have inspired extensive scrutiny regarding their effectivity in avoiding costly public rescues and bail-outs and potential adverse effects on market dynamics. While allowing banks to leverage and gain a higher return on their equity capital during the upturns in financial markets, contingent capital provides an automatic mechanism to reduce debt and raise the loss bearing capital cushion during the downturns and market crashes; therefore, making it possible to achieve stability and robustness in the financial sector, without reducing efficiency and competitiveness of the banking system with higher regulatory capital requirements. However, many researchers have raised concerns regarding unintended consequences and implications of such instruments for market dynamics. Death spirals in the stock price near the conversion, possibility of profitable stock or book manipulations by either the investors or the issuer, the marketability and demand for such hybrid instruments, contagion and systemic risks arising from the hedging strategies of the investors and higher risk taking incentives for issuers are among such concerns. Though substantial, many of such issues are addressed through a prudent design of the trigger and conversion mechanism. In the following chapters, we develop multiple models for pricing and analysis of contingent capital under different conversion mechanisms. In Chapter 2 we analyze the case of contingent capital with a capital-ratio trigger and partial and on-going conversion. The capital ratio we use is based on accounting or book value to approximate the regulatory ratios that determine capital requirements for banks. The conversion process is partial and on-going in the sense that each time a bank's capital ratio reaches the minimum threshold, just enough debt is converted to equity to meet the capital requirement, so long as the contingent capital has not been depleted. In Chapter 3 we simplify the design to all-at-once conversion however we perform the analysis through a much richer model which incorporates tail risk in terms of jumps, endogenous optimal default policy and debt rollover. We also investigate the case of bail-in debt, where at default the original shareholders are wiped out and the converted investors take control of the firm. In the case of contingent convertibles the conversion trigger is assumed as a contractual term specified by market value of assets. For bail-in debt the trigger is where the original shareholders optimally default. We study incentives of shareholders to change the capital structure and how CoCo's affect risk incentives. Several researchers have advocated use of a market based trigger which is forward looking, continuously updated and readily available, while some others have raised concerns regarding unintended consequences of a market based trigger. In Chapter 4 we investigate one of these issues, namely the existence and uniqueness of equilibrium when the conversion trigger is based on the stock price.Finance, Operations researchbn2164Industrial Engineering and Operations ResearchThesesThree Essays on Dynamic Pricing and Resource Allocation
https://academiccommons.columbia.edu/catalog/ac:151966
Nur, Cavdaroglu10.7916/D8W09D0GWed, 07 Jun 2017 17:00:14 +0000This thesis consists of three essays that focus on different aspects of pricing and resource allocation. We use techniques from supply chain and revenue management, scenario-based robust optimization and game theory to study the behavior of firms in different competitive and non-competitive settings. We develop dynamic programming models that account for pricing and resource allocation decisions of firms in such settings. In Chapter 2, we focus on the resource allocation problem of a service firm, particularly a health-care facility. We formulate a general model that is applicable to various resource allocation problems of a hospital. To this end, we consider a system with multiple customer classes that display different reactions to delays in service. By adopting a dynamic-programming approach, we show that the optimal policy is not simple but exhibits desirable monotonicity properties. Furthermore, we propose a simple threshold heuristic policy that performs well in our experiments. In Chapter 3, we study a dynamic pricing problem for a monopolist seller that operates in a setting where buyers have market power, and where each potential sale takes the form of a bilateral negotiation. We review the dynamic programming formulation of the negotiation problem, and propose a simple and tractable deterministic "fluid" analogue for this problem. The main emphasis of the chapter is in expanding the formulation to the dynamic setting where both the buyer and seller have limited prior information on their counterparty valuation and their negotiation skill. In Chapter 4, we consider the revenue maximization problem of a seller who operates in a market where there are two types of customers; namely the "investors" and "regular-buyers". In a two-period setting, we model and solve the pricing game between the seller and the investors in the latter period, and based on the solution of this game, we analyze the revenue maximization problem of the seller in the former period. Moreover, we study the effects on the total system profits when the seller and the investors cooperate through a contracting mechanism rather than competing with each other; and explore the contracting opportunities that lead to higher profits for both agents.Operations researchIndustrial Engineering and Operations ResearchThesesStrategic Models in Supply Network Design
https://academiccommons.columbia.edu/catalog/ac:147203
Lederman, Roger10.7916/D8X63V2VWed, 07 Jun 2017 16:58:54 +0000This dissertation contains a series of essays intended to introduce strategic modeling techniques into the network design problem. While investment in production capacity has long been approached as a critical strategic decision, the increasing need for robust, responsive supply capabilities has made it essential to take a network view, where multiple products and sites are considered simultaneously. In traditional network planning, models have rarely accounted for the behavior of additional players - customers, competitors, suppliers - on whom a firm can exert only a limited influence. We analyze a set of models that account for the dynamics of the firm's interaction with these outside actors. In Chapters 2 and 3, we develop game-theoretic models to characterize the allocation of resources in a network context. In Chapter 2, we use series-parallel networks to model the arrangement of producers whose output is bundled. This structure may arise, for example, when various components of the production process are outsourced individually. We study supply-function mechanisms through which producers strategically manage scarce capacity. Our results show how network structure can be analyzed to measure producers' market power and its effect on equilibrium markups. Chapter 3 looks at the network design problem of a vertically integrated firm with the ability to flexibly allocate resources across markets. We consider optimal design of the firm's production network as an upper-level decision to be optimized with respect to competitive outcomes in the lower stage. We find that optimal strategies regarding the location and centralization of production will differ across firms, depending on their competitive position in the market. The final two chapters discuss practical issues regarding the availability of model inputs in a multi-product context. In Chapter 4, we propose a method to construct competitor sets through estimation of a latent-segment choice model. We present a case study in a hotel market, where demand is distributed both spatially and temporally. We show how widely available data on market events can be used to drive identification of customer segments, providing a basis to assess competitive interactions. Chapter 5 provides a further example, in the setting of urban transportation networks, of how user behavior on a network can be estimated from partially observed data. We present a novel two-phase approach for performing this estimation in real time.Operations research, Businessrdl2102BusinessThesesModeling Customer Behavior for Revenue Management
https://academiccommons.columbia.edu/catalog/ac:151773
Bansal, Matulya10.7916/D8KP8867Wed, 07 Jun 2017 16:55:57 +0000In this thesis, we model and analyze the impact of two behavioral aspects of customer decision making upon the revenue maximization problem of a monopolist firm. First, we study the revenue maximization problem of a monopolist firm selling a homogeneous good to a market of risk-averse, strategic customers. Using a discrete (but arbitrary) valuation distribution, we show how the dynamic pricing problem with strategic customers can be formulated as a mechanism design problem, thereby making it more amenable to analysis. We characterize the optimal solution, and solve the problem for several special cases. We perform asymptotic analysis for the low risk-aversion case and show that it is asymptotically optimal to offer at most two products. Second, we consider a revenue-maximizing monopolist firm that serves a market of customers that are heterogeneous with respect to their valuations and desire for a quality attribute. Instead of optimizing the net utility that results from an appropriate combination of product price and quality, as in the traditional model of customer behavior, we consider a setting where customers purchase the cheapest product subject to its quality exceeding a customer specific quality threshold. We call such preferences threshold preferences. We solve the firm’s product design problem in this setting, and contrast with the traditional model of customer choice behavior. We consider several scenarios where such preferences might arise, and identify the optimal solution in each case. In addition to these product design problems, we study the problem of identifying the optimal putting strategy for a golfer. We develop a model of golfer putting skill, and combine it with a putt trajectory and holeout model to identify a golfer’s optimal putting strategy. The problem of identifying the optimal putting strategy is shown to be equivalent to a two-dimensional stochastic shortest path problem, with continuous state and control space, and solved using approximate dynamic programming. We calibrate the golfer model to professional and amateur player data, and use the calibrated model to answer several interesting questions, e.g., how does green reading ability affect golfer performance, how do professional and amateur golfers differ in their strategy, how do uphill and downhill putts compare in difficulty, etc.Business, Operations researchmb2431BusinessThesesEssays on Inventory Management and Object Allocation
https://academiccommons.columbia.edu/catalog/ac:144769
Lee, Thiam Hui10.7916/D8JM2HM7Wed, 07 Jun 2017 02:51:18 +0000This dissertation consists of three essays. In the first, we establish a framework for proving equivalences between mechanisms that allocate indivisible objects to agents. In the second, we study a newsvendor model where the inventory manager has access to two experts that provide advice, and examine how and when an optimal algorithm can be efficiently computed. In the third, we study classical single-resource capacity allocation problem and investigate the relationship between data availability and performance guarantees.
We first study mechanisms that solve the problem of allocating indivisible objects to agents. We consider the class of mechanisms that utilize the Top Trading Cycles (TTC) algorithm (these may differ based on how they prioritize agents), and show a general approach to proving equivalences between mechanisms from this class. This approach is used to show alternative and simpler proofs for two recent equivalence results for mechanisms with linear priority structures. We also use the same approach to show that these equivalence results can be generalized to mechanisms where the agent priority structure is described by a tree.
Second, we study the newsvendor model where the manager has recourse to advice, or decision recommendations, from two experts, and where the objective is to minimize worst-case regret from not following the advice of the better of the two agents. We show the model can be reduced to the class machine-learning problem of predicting binary sequences but with an asymmetric cost function, allowing us to obtain an optimal algorithm by modifying a well-known existing one. However, the algorithm we modify, and consequently the optimal algorithm we describe, is not known to be efficiently computable, because it requires evaluations of a function v which is the objective value of recursively defined optimization problems. We analyze v and show that when the two cost parameters of the newsvendor model are small multiples of a common factor, its evaluation is computationally efficient. We also provide a novel and direct asymptotic analysis of v that differs from previous approaches. Our asymptotic analysis gives us insight into the transient structure of v as its parameters scale, enabling us to formulate a heuristic for evaluating v generally. This, in turn, defines a heuristic for the optimal algorithm whose decisions we find in a numerical study to be close to optimal.
In our third essay, we study the classical single-resource capacity allocation problem. In particular, we analyze the relationship between data availability (in the form of demand samples) and performance guarantees for solutions derived from that data. This is done by describing a class of solutions called epsilon-backwards accurate policies and determining a suboptimality gap for this class of solutions. The suboptimality gap we find is in terms of epsilon and is also distribution-free. We then relate solutions generated by a Monte Carlo algorithm and epsilon-backwards accurate policies, showing a lower bound on the quantity of data necessary to ensure that the solution generated by the algorithm is epsilon-backwards accurate with a high probability. Combining the two results then allows us to give a lower bound on the data needed to generate an α-approximation with a given confidence probability 1-delta. We find that this lower bound is polynomial in the number of fares, M, and 1/α.Operations researchthl2102Industrial Engineering and Operations ResearchThesesEssays in Consumer Choice Driven Assortment Planning
https://academiccommons.columbia.edu/catalog/ac:131420
Saure, Denis R.10.7916/D8T4412PWed, 07 Jun 2017 02:51:03 +0000Product assortment selection is among the most critical decisions facing retailers: product variety and relevance is a fundamental driver of consumers' purchase decisions and ultimately of a retailer's profitability. In the last couple of decades an increasing number of firms have gained the ability to frequently revisit their assortment decisions during a selling season. In addition, the development and consolidation of online retailing have introduced new levels of operational flexibility, and cheap access to detailed transactional information. These new operational features present the retailer with both benefits and challenges. The ability to revisit the assortment decision frequently over time allows the retailer to introduce and test new products during the selling season, and adjust on the fly to unexpected changes in consumer preferences, and use customer profile information to customize (in real time) online shopping experience. Our main objective in this thesis is to formulate and solve assortment optimization models addressing the challenges present in modern retail environments. We begin by analyzing the role of the assortment decision in balancing information collection and revenue maximization, when consumer preferences are initially unknown. By considering utility maximizing consumers, we establish fundamental limits on the performance of any assortment policy whose aim is to maximize long run revenues. In addition, we propose adaptive assortment policies that attain such performance limits. Our results highlight salient features of this dynamic assortment problem that distinguish it from similar problems of sequential decision making under model uncertainty. Next, we extend the analysis to the case when additional consumer profile information is available; our primary motivation here is the emerging area of online advertisement. As in the previous setup, we identify fundamental performance limits and propose adaptive policies attaining these limits. Finally, we focus on the effects of competition and consumers' access to information on assortment strategies. In particular, we study competition among retailers when they have access to common products, i.e., products that are available to the competition, and where consumers have full information about the retailers' offerings. Our results shed light on equilibrium properties in such settings and the effect common products have on this behavior.Operations researchdrs2114BusinessThesesMany-Server Queues with Time-Varying Arrivals, Customer Abandonment, and non-Exponential Distributions
https://academiccommons.columbia.edu/catalog/ac:136569
Liu, Yunan10.7916/D8XW4RS9Wed, 07 Jun 2017 02:50:58 +0000This thesis develops deterministic heavy-traffic fluid approximations for many-server stochastic queueing models. The queueing models, with many homogeneous servers working independently in parallel, are intended to model large-scale service systems such as call centers and health care systems. Such models also have been employed to study communication, computing and manufacturing systems. The heavy-traffic approximations yield relatively simple formulas for quantities describing system performance, such as the expected number of customers waiting in the queue. The new performance approximations are valuable because, in the generality considered, these complex systems are not amenable to exact mathematical analysis. Since the approximate performance measures can be computed quite rapidly, they usefully complement more cumbersome computer simulation. Thus these heavy-traffic approximations can be used to improve capacity planning and operational control. More specifically, the heavy-traffic approximations here are for large-scale service systems, having many servers and a high arrival rate. The main focus is on systems that have time-varying arrival rates and staffing functions. The system is considered under the assumption that there are alternating periods of overloading and underloading, which commonly occurs when service providers are unable to adjust the staffing frequently enough to economically meet demand at all times. The models also allow the realistic features of customer abandonment and non-exponential probability distributions for the service times and the times customers are willing to wait before abandoning. These features make the overall stochastic model non-Markovian and thus thus very difficult to analyze directly. This thesis provides effective algorithms to compute approximate performance descriptions for these complex systems. These algorithms are based on ordinary differential equations and fixed point equations associated with contraction operators. Simulation experiments are conducted to verify that the approximations are effective. This thesis consists of four pieces of work, each presented in one chapter. The first chapter (Chapter 2) develops the basic fluid approximation for a non-Markovian many-server queue with time-varying arrival rate and staffing. The second chapter (Chapter 3) extends the fluid approximation to systems with complex network structure and Markovian routing to other queues of customers after completing service from each queue. The extension to open networks of queues has important applications. For one example, in hospitals, patients usually move among different units such as emergency rooms, operating rooms, and intensive care units. For another example, in manufacturing systems, individual products visit different work stations one or more times. The open network fluid model has multiple queues each of which has a time-varying arrival rate and staffing function. The third chapter (Chapter 4) studies the large-time asymptotic dynamics of a single fluid queue. When the model parameters are constant, convergence to the steady state as time evolves is established. When the arrival rates are periodic functions, such as in service systems with daily or seasonal cycles, the existence of a periodic steady state and the convergence to that periodic steady state as time evolves are established. Conditions are provided under which this convergence is exponentially fast. The fourth chapter (Chapter 5) uses a fluid approximation to gain insight into nearly periodic behavior seen in overloaded stationary many-server queues with customer abandonment and nearly deterministic service times. Deterministic service times are of applied interest because computer-generated service times, such as automated messages, may well be deterministic, and computer-generated service is becoming more prevalent. With deterministic service times, if all the servers remain busy for a long interval of time, then the times customers enter service assumes a periodic behavior throughout that interval. In overloaded large-scale systems, these intervals tend to persist for a long time, producing nearly periodic behavior. To gain insight, a heavy-traffic limit theorem is established showing that the fluid model arises as the many-server heavy-traffic limit of a sequence of appropriately scaled queueing models, all having these deterministic service times. Simulation experiments confirm that the transient behavior of the limiting fluid model provides a useful description of the transient performance of the queueing system. However, unlike the asymptotic loss of memory results in the previous chapter for service times with densities, the stationary fluid model with deterministic service times does not approach steady state as time evolves independent of the initial conditions. Since the queueing model with deterministic service times approaches a proper steady state as time evolves, this model with deterministic service times provides an example where the limit interchange (limiting steady state as time evolves and heavy traffic as scale increases) is not valid.Operations researchyl2342Industrial Engineering and Operations ResearchThesesAlgorithms for Sparse and Low-Rank Optimization: Convergence, Complexity and Applications
https://academiccommons.columbia.edu/catalog/ac:137539
Ma, Shiqian10.7916/D8PC38BZWed, 07 Jun 2017 02:50:57 +0000Solving optimization problems with sparse or low-rank optimal solutions has been an important topic since the recent emergence of compressed sensing and its matrix extensions such as the matrix rank minimization and robust principal component analysis problems. Compressed sensing enables one to recover a signal or image with fewer observations than the "length" of the signal or image, and thus provides potential breakthroughs in applications where data acquisition is costly. However, the potential impact of compressed sensing cannot be realized without efficient optimization algorithms that can handle extremely large-scale and dense data from real applications. Although the convex relaxations of these problems can be reformulated as either linear programming, second-order cone programming or semidefinite programming problems, the standard methods for solving these relaxations are not applicable because the problems are usually of huge size and contain dense data. In this dissertation, we give efficient algorithms for solving these "sparse" optimization problems and analyze the convergence and iteration complexity properties of these algorithms. Chapter 2 presents algorithms for solving the linearly constrained matrix rank minimization problem. The tightest convex relaxation of this problem is the linearly constrained nuclear norm minimization. Although the latter can be cast and solved as a semidefinite programming problem, such an approach is computationally expensive when the matrices are large. In Chapter 2, we propose fixed-point and Bregman iterative algorithms for solving the nuclear norm minimization problem and prove convergence of the first of these algorithms. By using a homotopy approach together with an approximate singular value decomposition procedure, we get a very fast, robust and powerful algorithm, which we call FPCA (Fixed Point Continuation with Approximate SVD), that can solve very large matrix rank minimization problems. Our numerical results on randomly generated and real matrix completion problems demonstrate that this algorithm is much faster and provides much better recoverability than semidefinite programming solvers such as SDPT3. For example, our algorithm can recover 1000 × 1000 matrices of rank 50 with a relative error of 10<sup>-5</sup> in about 3 minutes by sampling only 20 percent of the elements. We know of no other method that achieves as good recoverability. Numerical experiments on online recommendation, DNA microarray data set and image inpainting problems demonstrate the effectiveness of our algorithms. In Chapter 3, we study the convergence/recoverability properties of the fixed point continuation algorithm and its variants for matrix rank minimization. Heuristics for determining the rank of the matrix when its true rank is not known are also proposed. Some of these algorithms are closely related to greedy algorithms in compressed sensing. Numerical results for these algorithms for solving linearly constrained matrix rank minimization problems are reported. Chapters 4 and 5 considers alternating direction type methods for solving composite convex optimization problems. We present in Chapter 4 alternating linearization algorithms that are based on an alternating direction augmented Lagrangian approach for minimizing the sum of two convex functions. Our basic methods require at most O(1/ε) iterations to obtain an ε-optimal solution, while our accelerated (i.e., fast) versions require at most O(1/√ε) iterations, with little change in the computational effort required at each iteration. For more general problem, i.e., minimizing the sum of K convex functions, we propose multiple-splitting algorithms for solving them. We propose both basic and accelerated algorithms with O(1/ε) and O(1/√ε) iteration complexity bounds for obtaining an ε-optimal solution. To the best of our knowledge, the complexity results presented in these two chapters are the first ones of this type that have been given for splitting and alternating direction type methods. Numerical results on various applications in sparse and low-rank optimization, including compressed sensing, matrix completion, image deblurring, robust principal component analysis, are reported to demonstrate the efficiency of our methods.Operations researchsm2756Industrial Engineering and Operations ResearchThesesSupply Chain Management: Supplier Financing Schemes and Inventory Strategies
https://academiccommons.columbia.edu/catalog/ac:142635
Wang, Min10.7916/D81R6XGNWed, 07 Jun 2017 02:46:26 +0000This dissertation addresses a few fundamental questions on the interface between supplier financing schemes and inventory management. Traditionally, retailers finance their inventories through an independent financing institution or by drawing from their own cash reserves, without any supplier involvement (Independent Financing). However, suppliers may reduce their buyers' costs and stimulate sales and associated revenues and profits, by either (i) adopting the financing function themselves (Trade Credit), or (ii) subsidizing the inventory costs (Inventory Subsidies). In the first part (Chapter 2) we analyze and compare the equilibrium performance of supply chains under these three basic financing schemes. The objective is to compare the equilibrium profits of the individual chain members, the aggregate supply chain profits, the equilibrium wholesale price, the expected sales volumes and the average inventory levels under the three financing options, and thus provide important insights for the selection and implementation of supply chain financing mechanisms. Several of the financing schemes introduce a new type of inventory control problem for the retailers in response to terms specified by their suppliers. In Chapter 3 we therefore consider the inventory management problem of a firm which incurs inventory carrying costs with a general shelf age dependent structure and, even more generally, that of a firm with shelf age and delay dependent inventory and backlogging costs. Beyond identifying the structure of optimal replenishment strategies and corresponding algorithms to compute them, it is often important to understand how changes in various primitives of the inventory model impact on the optimal policy parameters and performance measures. In spite of a voluminous literature over more than fifty years, very little is known about this area. In Chapter 4, we therefore study monotonicity properties of stochastic inventory systems governed by an (r; q) or (r; nq) policy and apply the results in our general theorems both to standard inventory models and to those with general shelf age and delay dependent inventory costs.Business, Operations researchmw2426BusinessThesesA Simulation Model to Analyze the Impact of Golf Skills and a Scenario-based Approach to Options Portfolio Optimization
https://academiccommons.columbia.edu/catalog/ac:143076
Ko, Soonmin10.7916/D85D8ZT3Wed, 07 Jun 2017 02:46:24 +0000A simulation model of the game of golf is developed to analyze the impact of various skills (e.g., driving distance, directional accuracy, putting skill, and others) on golf scores. The golf course model includes realistic features of a golf course including rough, sand, water, and trees. Golfer shot patterns are modeled with t distributions and mixtures of t and normal distributions since normal distributions do not provide good fits to the data. The model is calibrated to extensive data for amateur and professional golfers. The golf simulation is used to assess the impact on scores of distance and direction, determine what factors separate pros from amateurs, and to determine the impact of course length on scores. In the second part of the thesis, we use a scenario-based approach to solve a portfolio optimization problem with options. The solution provides the optimal payoff profile given an investor's view of the future, his utility function or risk appetite, and the market prices of options. The scenario-based approach has several advantages over the traditional covariance matrix method, including additional flexibility in the choice of constraints and objective function.Engineering, Operations researchsk2822Industrial Engineering and Operations ResearchThesesFirst Order Methods for Large-Scale Sparse Optimization
https://academiccommons.columbia.edu/catalog/ac:135750
Aybat, Necdet Serhat10.7916/D8V69RJJWed, 07 Jun 2017 02:46:03 +0000In today's digital world, improvements in acquisition and storage technology are allowing us to acquire more accurate and finer application-specific data, whether it be tick-by-tick price data from the stock market or frame-by-frame high resolution images and videos from surveillance systems, remote sensing satellites and biomedical imaging systems. Many important large-scale applications can be modeled as optimization problems with millions of decision variables. Very often, the desired solution is sparse in some form, either because the optimal solution is indeed sparse, or because a sparse solution has some desirable properties. Sparse and low-rank solutions to large scale optimization problems are typically obtained by regularizing the objective function with L1 and nuclear norms, respectively. Practical instances of these problems are very high dimensional (~ million variables) and typically have dense and ill-conditioned data matrices. Therefore, interior point based methods are ill-suited for solving these problems. The large scale of these problems forces one to use the so-called first-order methods that only use gradient information at each iterate. These methods are efficient for problems with a "simple" feasible set such that Euclidean projections onto the set can be computed very efficiently, e.g. the positive orthant, the n-dimensional hypercube, the simplex, and the Euclidean ball. When the feasible set is "simple", the subproblems used to compute the iterates can be solved efficiently. Unfortunately, most applications do not have "simple" feasible sets. A commonly used technique to handle general constraints is to relax them so that the resulting problem has only "simple" constraints, and then to solve a single penalty or Lagrangian problem. However, these methods generally do not guarantee convergence to feasibility. The focus of this thesis is on developing new fast first-order iterative algorithms for computing sparse and low-rank solutions to large-scale optimization problems with very mild restrictions on the feasible set - we allow linear equalities, norm-ball and conic inequalities, and also certain non-smooth convex inequalities to define the constraint set. The proposed algorithms guarantee that the sequence of iterates converges to an optimal feasible solution of the original problem, and each subproblem is an optimization problem with a "simple" feasible set. In addition, for any eps > 0, by relaxing the feasibility requirement of each iteration, the proposed algorithms can compute an eps-optimal and eps-feasible solution within O(log(1/eps)) iterations which requires O(1/eps) basic operations in the worst case. Algorithm parameters do not depend on eps > 0. Thus, these new methods compute iterates arbitrarily close to feasibility and optimality as they continue to run. Moreover, the computational complexity of each basic operation for these new algorithms is the same as that of existing first-order algorithms running on "simple" feasible sets. Our numerical studies showed that only O(log(1/eps)) basic operations, as opposed to O(1/eps) worst case theoretical bound, are needed for obtaining eps-feasible and eps-optimal solutions. We have implemented these new first-order methods for the following problem classes: Basis Pursuit (BP) in compressed sensing, Matrix Rank Minimization, Principal Component Pursuit (PCP) and Stable Principal Component Pursuit (SPCP) in principal component analysis. These problems have applications in signal and image processing, video surveillance, face recognition, latent semantic indexing, and ranking and collaborative filtering. To best of our knowledge, an algorithm for the SPCP problem that has O(1/eps) iteration complexity and has a per iteration complexity equal to that of a singular value decomposition is given for the first time.Operations research, Mathematicsnsa2106Industrial Engineering and Operations ResearchThesesMultiproduct Pricing Management and Design of New Service Products
https://academiccommons.columbia.edu/catalog/ac:144706
Wang, Ruxian10.7916/D81V5MXDWed, 07 Jun 2017 02:45:51 +0000In this thesis, we study price optimization and competition of multiple differentiated substitutable products under the general Nested Logit model and also consider the designing and pricing of new service products, e.g., flexible warranty and refundable warranty, under customers' strategic claim behavior. Chapter 2 considers firms that sell multiple differentiated substitutable products and customers whose purchase behavior follows the Nested Logit model, of which the Multinomial Logit model is a special case. In the Nested Logit model, customers make product selection decision sequentially: they first select a class or a nest of products and subsequently choose a product within the selected class. We consider the general Nested Logit model with product-differentiated price coefficients and general nest-heterogenous degrees. We show that the adjusted markup, which is defined as price minus cost minus the reciprocal of the price coefficient, is constant across all the products in each nest. When optimizing multiple nests of products, the adjusted nested markup is also constant within a nest. By using this result, the multi-product optimization problem can be reduced to a single-dimensional problem in a bounded interval, which is easy to solve. We also use this result to simplify the oligopolistic price competition and characterize the Nash equilibrium. Furthermore, we investigate its application to dynamic pricing and revenue management. In Chapter 3, we investigate the flexible monthly warranty, which offers flexibility to customers and allow them to cancel it at anytime without any penalty. Frequent technological innovations and price declines severely affect sales of extended warranties as product replacement upon failure becomes an increasingly attractive alternative. To increase sales and profitability, we propose offering flexible-duration extended warranties. These warranties can appeal to customers who are uncertain about how long they will keep the product as well as to customers who are uncertain about the product's reliability. Flexibility may be added to existing services in the form of monthly-billing with month-by-month commitments, or by making existing warranties easier to cancel, with pro-rated refunds. This thesis studies flexible warranties from the perspectives of both the customer and the provider. We present a model of the customer's optimal coverage decisions under the objective of minimizing expected support costs over a random planning horizon. We show that under some mild conditions the customer's optimal coverage policy has a threshold structure. We also show through an analytical study and through numerical examples how flexible warranties can result in higher profits and higher attach rates. Chapter 4 examines the designing and pricing of residual value warranty that refunds customers at the end of warranty period based on customers' claim history. Traditional extended warranties for IT products do not differentiate customers according to their usage rates or operating environment. These warranties are priced to cover the costs of high-usage customers who tend to experience more failures and are therefore more costly to support. This makes traditional warranties economically unattractive to low-usage customers. In this chapter, we introduce, design and price residual value warranties. These warranties refund a part of the upfront price to customers who have zero or few claims according to a pre-determined refund schedule. By design, the net cost of these warranties is lower for light users than for heavy users. As a result, a residual value warranty can enable the provider to price-discriminate based on usage rates or operating conditions without the need to monitor individual customers' usage. Theoretic results and numerical experiments demonstrate how residual value warranties can appeal to a broader range of customers and significantly increase the provider's profits.Operations research, Industrial engineeringrw2267Industrial Engineering and Operations ResearchThesesStability and Structural Properties of Stochastic Storage Networks
https://academiccommons.columbia.edu/catalog/ac:v15dv41nv6
Kella, Offer; Whitt, Ward10.7916/D8TM7PK7Thu, 25 May 2017 18:46:41 +0000We establish stability, monotonicity, concavity and subadditivity properties for open stochastic storage networks in which the driving process has stationary increments. A principal example is a stochastic fluid network in which the external inputs are random but all internal flows are deterministic. For the general model, the multi-dimensional content process is tight under the natural stability condition. The multi-dimensional content process is also stochastically increasing when the process starts at the origin, implying convergence to a proper limit under the natural stability condition. In addition, the content process is monotone in its initial conditions. Hence, when any content process with nonzero initial conditions hits the origin, it couples with the content process starting at the origin. However, in general, a tight content process need not hit the origin.Stochastic systems, Lévy processes, Stochastic systems, Industrial engineering, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesMarkov Chain Models to Estimate the Premium for Extended Hedge Fund Lockups
https://academiccommons.columbia.edu/catalog/ac:9cnp5hqc0g
Derman, Emanuel; Park, Kun Soo; Whitt, Ward10.7916/D8BP0F80Thu, 25 May 2017 18:37:04 +0000A lockup period for investment in a hedge fund is a time period after making the investment during which the investor cannot freely redeem his investment. It is routine to have a one-year lockup period, but recently the requested lockup periods have grown longer. We estimate the premium for such extended lockup, taking the point of view of a manager of a fund of funds, who has to choose between two investments in similar funds in the same strategy category, the first having a one-year lockup and the second having an n-year lockup. Assuming that the manager will rebalance his portfolio of hedge funds on a yearly basis, if permitted, we define the annual lockup premium as the difference between the rates of return from these investments. We develop a Markov chain model to estimate this lockup premium. By solving systems of equations, we fit the Markov chain transition probabilities to three directly observable hedge fund performance measures: the persistence of return, the variance of return and the hedge-fund death rate. The model quantifies the way the lockup premium depends on these parameters. Data from the TASS database are used to estimate the persistence, which is found to be statistically significant.Hedge funds, Liquidity (Economics), Markov processes, Stochastic systems, Industrial engineering, Probabilities, Queuing theory, Operations researched2114, ww2040Industrial Engineering and Operations ResearchArticlesWhat you should know about queueing models to set staffing requirements in service systems
https://academiccommons.columbia.edu/catalog/ac:fqz612jm7c
Whitt, Ward10.7916/D88P6C1PThu, 25 May 2017 18:37:02 +0000One traditional application of queueing models is to help set staffing requirements in service systems, but the way to do so is not entirely straightforward, largely because demand in service systems typically varies greatly by the time of day. This article discusses ways - old and new - to cope with that time-varying demand.Call centers, Queuing theory, Queuing theory, Queuing networks (Data transmission), Industrial engineering, Probabilities, Operations researchww2040Industrial Engineering and Operations ResearchArticlesWorkload bounds in fluid models with priorities
https://academiccommons.columbia.edu/catalog/ac:dbrv15dv54
Berger, Arthur W.; Whitt, Ward10.7916/D8RV116ZThu, 25 May 2017 18:36:56 +0000In this paper we establish upper and lower bounds on the steady-state per-class workload distributions in a single-server queue with multiple priority classes. Motivated by communication network applications, the model has constant processing rate and general input processes with stationary increments. The bounds involve corresponding quantities in related models with the first-come first-served discipline. We apply the bounds to support a new notion of effective bandwidths for multi-class systems with priorities. We also apply the lower bound to obtain sufficient conditions for the workload distributions to have heavy tails.Queuing theory, Stochastic models, Industrial engineering, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesLinear stochastic fluid networks
https://academiccommons.columbia.edu/catalog/ac:d51c59zw58
Kella, Offer; Whitt, Ward10.7916/D89P3D52Thu, 25 May 2017 18:36:47 +0000We introduce open stochastic fluid networks that can be regarded as continuous analogues or fluid limits of open networks of infinite-server queues. Random exogenous input may come to any of the queues. At each queue, a c.d.f.-valued stochastic process governs the proportion of the input processed by a given time after arrival. The routeing may be deterministic (a specified sequence of successive queue visits) or proportional, i.e. a stochastic transition matrix may govern the proportion of the output routed from one queue to another. This stochastic fluid network with deterministic c.d.f.s governing processing at the queues arises as the limit of normalized networks of infinite-server queues with batch arrival processes where the batch sizes grow. In this limit, one can think of each particle having an evolution through the network, depending on its time and place of arrival, but otherwise independent of all other particles. A key property associated with this independence is the linearity: the workload associated with a superposition of inputs, each possibly having its own pattern of flow through the network, is simply the sum of the component workloads. As with infinite-server queueing models, the tractability makes the linear stochastic fluid network a natural candidate for approximations.Stochastic systems, Storage area networks (Computer networks), Lévy processes, Queuing networks (Data transmission), Industrial engineering, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesDecomposition approximations for time-dependent Markovian queueing networks
https://academiccommons.columbia.edu/catalog/ac:r7sqv9s4pp
Whitt, Ward10.7916/D8H99HPCThu, 25 May 2017 18:36:41 +0000Motivated by the development of complex telephone call center networks, we present a general framework for decompositions to approximately solve Markovian queueing networks with time-dependent and state-dependent transition rates. The decompositions are based on assuming either full or partial product form for the time-dependent probability vectors at each time. These decompositions reduce the number of time-dependent ordinary differential equations that must be solved. We show how special structure in the transition rates can be exploited to speed up computation. There is extra theoretical support for the decomposition approximation when the steady-state distribution of the time-homogeneous version of the model has product form.Queuing theory, Queuing networks (Data transmission), Markov processes, Call centers, Industrial engineering, Probabilities, Operations researchww2040Industrial Engineering and Operations ResearchArticlesUsing different response-time requirements to smooth time-varying demand for service
https://academiccommons.columbia.edu/catalog/ac:02v6wwpzgw
Whitt, Ward10.7916/D87W6QNNThu, 25 May 2017 18:36:38 +0000Many service systems have demand that varies significantly by time of day, making it costly to provide sufficient capacity to be able to respond very quickly to each service re- quest. Fortunately, however, different service requests often have very different response-time requirements. Some service requests may need immediate response, while others can tolerate substantial delays. Thus it is often possible to smooth demand by partitioning the service requests into separate priority classes according to their response-time requirements. Classes with more stringent performance requirements are given higher priority for service. Lower capacity may be required if lower-priority-class demand can be met during off-peak periods. We show how the priority classes can be defined and the resulting required fixed capacity can be determined, directly accounting for the time-dependent behavior. For this purpose, we ex- ploit relatively simple analytical models, in particular, Mt/G/∞ and deterministic offered-load models. The analysis also provides an estimate of the capacity savings that can be obtained from partitioning time-varying demand into priority classes.Queuing theory, Queuing networks (Data transmission), Industrial engineering, Probabilities, Operations researchww2040Industrial Engineering and Operations ResearchArticlesAnalysis of join-the-shortest-queue routing for web server farms
https://academiccommons.columbia.edu/catalog/ac:xksn02v700
Gupta, Varun; Harchol-Balter, Mor; Sigman, Karl; Whitt, Ward10.7916/D81J9P84Thu, 25 May 2017 18:36:35 +0000Join the Shortest Queue (JSQ) is a popular routing policy for server farms. However, until now all analysis of JSQ has been limited to First-Come-First-Serve (FCFS) server farms, whereas it is known that web server farms are better modeled as Processor Sharing (PS) server farms. We provide the first approximate analysis of JSQ in the PS server farm model for general job-size distributions, obtaining the distribution of queue length at each queue. To do this, we approximate the queue length of each queue in the server farm by a one-dimensional Markov chain, in a novel fashion. We also discover some interesting insensitivity properties of PS server farms with JSQ routing, and discuss the near-optimality of JSQ.Queuing theory, Industrial engineering, Probabilities, Operations researchvg2297, ks20, ww2040Industrial Engineering and Operations ResearchArticlesCoping with Time-Varying Demand When Setting Staffing Requirements for a Service System
https://academiccommons.columbia.edu/catalog/ac:gmsbcc2fsb
Green, Linda V.; Kolesar, Peter J.; Whitt, Ward10.7916/D8RR29QBThu, 25 May 2017 18:36:32 +0000We review queueing-theory methods for setting staffing requirements in service systems where customer demand varies in a predictable pattern over the day. Analyzing these systems is not straightforward, because standard queueing theory focuses on the long-run steady-state behavior of stationary models. We show how to adapt stationary queueing models for use in nonstationary environments so that time-dependent performance is captured and staffing requirements can be set. Relatively little modification of straightforward stationary analysis applies in systems where service times are short and the targeted quality of service is high. When service times are moderate and the targeted quality of service is still high, time-lag refinements can improve traditional stationary independent period-by-period and peak-hour approximations. Time-varying infinite-server models help develop refinements, because closed-form expressions exist for their time-dependent behavior. More difficult cases with very long service times and other complicated features, such as end-of-day effects, can often be treated by a modified-offered-load approximation, which is based on an associated infinite-server model. Numerical algorithms and deterministic fluid models are useful when the system is overloaded for an extensive period of time. Our discussion focuses on telephone call centers, but applications to police patrol, banking, and hospital emergency rooms are also mentioned.Call centers, Queuing theory, Police patrol--Mathematical models, Banks and banking, Hospitals--Emergency services, Industrial engineering, Probabilities, Queuing theory, Operations researchlvg1, pjk4, ww2040Industrial Engineering and Operations ResearchArticlesValue-Based Routing and Preference-Based Routing in Customer Contact Centers
https://academiccommons.columbia.edu/catalog/ac:v6wwpzgmvb
Sisselman, Michael E.; Whitt, Ward10.7916/D8H70T92Thu, 25 May 2017 18:36:28 +0000Telephone call centers and their generalizations—customer contact centers—usually handle several types of customer service requests (calls). Since customer service representatives (agents) have different call-handling abilities and are typically cross-trained in multiple skills, contact centers exploit skill-based routing (SBR) to assign calls to appropriate agents, aiming to respond properly as well as promptly. Established agent-staffing and SBR algorithms ensure that agents have the required call-handling skills and that call routing is performed so that constraints are met for standard congestion measures, such as the percentage of calls of each type that abandon before starting service and the percentage of answered calls of each type that are delayed more than a specified number of seconds. We propose going beyond traditional congestion measures to focus on the expected value derived from having particular agents handle various calls. Expected value might represent expected revenue or the likelihood of first-call resolution. Value might also reflect agent call-handling preferences. We show how value-based routing (VBR) and preference-based routing (PBR) can be introduced in the context of an existing SBR framework, based on static-priority routing using a highly-structured priority matrix, so that constraints are still met for standard congestion measures. Since an existing SBR framework is used to implement VBR and PBR, it is not necessary to replace the automatic call distributor (ACD). We show how mathematical programming can be used, with established staffing requirements, to find a desirable priority matrix. We select the priority matrix to use during a specified time interval (e.g., 30-minute period) by maximizing the total expected value over that time interval, subject to staffing constraints.Customer services, Call centers, Labor turnover, Industrial engineering, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesHeavy-traffic extreme-value limits for Erlang delay models
https://academiccommons.columbia.edu/catalog/ac:jdfn2z34w1
Pang, Guodong; Whitt, Ward10.7916/D8251WQXThu, 25 May 2017 18:36:24 +0000We consider the maximum queue length and the maximum number of idle servers in the classical Erlang delay model and the generalization allowing customer abandonment – the M/M/n + M queue. We use strong approximations to show, under regularity conditions, that properly scaled versions of the maximum queue length and maximum number of idle servers over subintervals [0, t] in the delay models converge jointly to independent random variables with the Gumbel extreme- value distribution in the QED and ED many-server heavy-traffic limiting regimes as n and t increase to infinity together appropriately; we require that tn → ∞ and tn = o(n1/2−ε) as n → ∞ for some ε>0.Queuing theory, Queuing networks (Data transmission), Limit theorems (Probability theory), Industrial engineering, Probabilities, Operations researchgp2224, ww2040Industrial Engineering and Operations ResearchArticlesExplicit M/G/1 waiting-time distributions for a class of long-tail service-time distributions
https://academiccommons.columbia.edu/catalog/ac:stqjq2bvs8
Abate, Joseph; Whitt, Ward10.7916/D8WM1RXWThu, 25 May 2017 18:36:22 +0000O. J. Boxma and J. W. Cohen recently obtained an explicit expression for the M/G/1 steady-state waiting-time distribution for a class of service-time distributions with power tails. We extend their explicit representation from a one-parameter family of service-time distributions to a two-parameter family. The complementary cumulative distribution function (ccdf’s) of the service times all have the asymptotic form Fc(t) ∼ αt−3/2 as t → ∞, so that the associated waiting-time ccdf’s have asymptotic form Wc(t) ∼ βt−1/2 as t → ∞. Thus the second moment of the service time and the mean of the waiting time are infinite. Our result here also extends our own earlier explicit expression for the M/G/1 steady-state waiting-time distribution when the service-time distribution is an exponential mixture of inverse Gaussian distributions (EMIG). The EMIG distributions form a two-parameter family with ccdf hav- ing the asymptotic form Fc(t) ∼ αt−3/2e−ηt as t → ∞. We now show that a variant of our previous argument applies when the service-time ccdf is an undamped EMIG, i.e., with ccdf Gc(t) = eηtFc(t) for Fc(t) above, which has the power tail Gc(t) ∼ αt−3/2 as t → ∞. The Boxma-Cohen long-tail service-time distribution is a special case of an undamped EMIG.Queuing theory, Inverse Gaussian distribution, Industrial engineering, Probabilities, Operations researchww2040Industrial Engineering and Operations ResearchArticlesStaffing a Call Center with Uncertain Arrival Rate and Absenteeism
https://academiccommons.columbia.edu/catalog/ac:m905qfttgm
Whitt, Ward10.7916/D8M3377TThu, 25 May 2017 18:36:18 +0000This paper proposes simple methods for staffing a single-class call center with uncertain arrival rate and uncertain staffing due to employee absenteeism. The arrival rate and the proportion of servers present are treated as random variables. The basic model is a multi-server queue with customer abandonment, allowing non-exponential service-time and time-to-abandon distributions. The goal is to maximize the expected net return, given throughput benefit and server, customer-abandonment and customer-waiting costs, but attention is also given to the standard deviation of the return. The approach is to approximate the performance and the net return, conditional on the random model-parameter vector, and then uncondition to get the desired results. Two recently-developed approximations are used for the conditional performance measures: first, a deterministic fluid approximation and, second, a numerical algorithm based on a purely Markovian birth-and-death model, having state-dependent death rates.Uncertainty (Information theory), Call centers, Absenteeism (Labor), Industrial engineering, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesHeavy-traffic extreme-value limits for queues
https://academiccommons.columbia.edu/catalog/ac:jsxksn02wr
Glynn, Peter W.; Whitt, Ward10.7916/D80C5786Thu, 25 May 2017 18:36:16 +0000We consider the maximum waiting time among the first n customers in the GI/G/1 queue. We use strong approximations to prove, under regularity conditions, convergence of the normalized maximum wait to the Gumbel extreme-value distribution when the traffic intensity ρ approaches 1 from below and n approaches infinity at a suitable rate. The normalization depends on the interarrival-time and service-time distributions only through their first two moments, corresponding to the iterated limit in which first ρ approaches 1 and then n approaches infinity. We need n to approach infinity sufficiently fast so that n ( 1 − ρ )2 → ∞. We also need n to approach infinity sufficiently slowly: If the service time has a pth moment for ρ > 2, then it suffices for ( 1 − ρ ) n 1 / p to remain bounded; if the service time has a finite moment generating function, then it suffices to have ( 1 − ρ ) log n → 0. This limit can hold even when the normalized maximum waiting time fails to converge to the Gumbel distribution as n → ∞ for each fixed ρ. Similar limits hold for the queue length process.Queuing theory, Limit theorems (Probability theory), Industrial engineering, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesLimits for Cumulative Input Processes to Queues
https://academiccommons.columbia.edu/catalog/ac:05qfttdz09
Whitt, Ward10.7916/D8348XW6Thu, 25 May 2017 18:36:13 +0000We establish functional central limit theorems (FCLTs) for a cumulative input process to a fluid queue from the superposition of independent on—off sources, where the on periods and off periods may have heavy-tailed probability distributions. Variants of these FCLTs hold for cumulative busy-time and idle-time processes associated with standard queueing models. The heavy-tailed on-period and off-period distributions can cause the limit process to have discontinuous sample paths (e.g., to be a non-Brownian stable process or more general Lévy process) even though the converging processes have continuous sample paths. Consequently, we exploit the Skorohod M1 topology on the function space D of right-continuous functions with left limits. The limits here combined with the previously established continuity of the reflection map in the M1 topology imply both heavy-traffic and non-heavy-traffic FCLTs for buffer-content processes in stochastic fluid networks.Central limit theorem, Symmetry (Mathematics), Industrial engineering, Stochastic systems, Lévy processes, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesDynamic staffing in a telephone call center aiming to immediately answer all calls
https://academiccommons.columbia.edu/catalog/ac:7h44j0zpd2
Whitt, Ward10.7916/D8J67VDFThu, 25 May 2017 18:36:10 +0000This paper proposes practical modeling and analysis methods to facilitate dynamic staffing in a telephone call center with the objective of immediately answering all calls. Because of this goal, it is natural to use infinite-server queueing models. These models are very useful because they are so tractable. A key to the dynamic staffing is exploiting detailed knowledge of system state in order to obtain good estimates of the mean and variance of the demand in the near future. The near-term staffing needs, e.g., for the next minute or the next 20 min., can often be predicted by exploiting information about recent demand and current calls in progress, as well as historical data. The remaining holding times of calls in progress can be predicted by classifying and keeping track of call types, by measuring holding-time distributions and by taking account of the elapsed holding times of calls in progress. The number of new calls in service can be predicted by exploiting information about both historical and recent demand.Call centers, Forecasting, Queuing networks (Data transmission), Call centers, Industrial engineering, Probabilities, Queuing theory, Operations researchww2040Industrial Engineering and Operations ResearchArticlesHeavy-traffic limits for many-server queues with service interruptions
https://academiccommons.columbia.edu/catalog/ac:d51c59zw57
Pang, Guodong; Whitt, Ward10.7916/D8SN0NGVThu, 25 May 2017 18:36:07 +0000We establish many-server heavy-traffic limits for G/M/n + M queueing models, allowing cus- tomer abandonment (the +M), subject to exogenous regenerative service interruptions. With unscaled service interruption times, we obtain a FWLLN for the queue-length process, where the limit is an ordinary differential equation in a two-state random environment. With asymptoti- cally negligible service interruptions, we obtain a FCLT for the queue-length process, where the limit is characterized as the pathwise unique solution to a stochastic integral equation with jumps. When the arrivals are renewal and the interruption cycle time is exponential, the limit is a Markov process, being a jump-diffusion process in the QED regime and an O-U process driven by a Levy process in the ED regime (and for infinite-server queues). A stochastic-decompostion property of the steady-state distribution of the limit process in the ED regime (and for infinite-server queues) is obtained.Queuing networks (Data transmission), Lévy processes, Industrial engineering, Probabilities, Queuing theory, Operations researchgp2224, ww2040Industrial Engineering and Operations ResearchArticlesTwo fluid approximations for multi-server queues with abandonments
https://academiccommons.columbia.edu/catalog/ac:x3ffbg79fp
Whitt, Ward10.7916/D8154VHNThu, 25 May 2017 18:36:05 +0000Insight is provided into a previously developed M/M/s/r+M(n) approximation for the M/GI/s/r+GI queueing model by establishing fluid and diffusion limits for the approximating model. Fluid approximations for the two models are compared in the many-server efficiency-driven (overloaded) regime. The two fluid approximations do not coincide, but they are close.Queuing networks (Data transmission), Queuing theory, Call centers, Industrial engineering, Probabilities, Operations researchww2040Industrial Engineering and Operations ResearchArticles