\documentclass[twoside, 12pt]{article} \usepackage{mathptmx} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \usepackage{latexsym} \usepackage{amsmath,amsthm,amssymb} \usepackage{epsfig} \usepackage{stmaryrd} \usepackage{multicol} \usepackage{pdfsync} \usepackage{dsfont} \usepackage{comment} %\usepackage{natbib} \usepackage{caption} \usepackage{psfrag} \usepackage[final=true,colorlinks=true,breaklinks=true,linkcolor=black,citecolor=black]{hyperref} %\usepackage{subcaption} %\usepackage{subfig} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %\usepackage{mor} \usepackage{graphics} %\usepackage[nohead, margin=1.0in, truedimen]{geometry} \DeclareGraphicsExtensions{.png,.pdf} %%%%%%%%%%%%%% % page size \setlength{\hoffset}{0pt} \setlength{\voffset}{0pt} \setlength{\headsep}{-30pt} \setlength{\headheight}{0pt} \setlength{\footskip}{30pt} \setlength{\textwidth}{7.0 truein} \setlength{\textheight}{9.8 truein} \setlength{\topmargin}{0 truein} \setlength{\oddsidemargin}{-0.2in} \setlength{\evensidemargin}{-0.2in} \pagestyle{empty} % no numbering %%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%% ---- Author's inserts ----%%%%%%%%%%% \begin{document} %\title{Research Statement} %\author{Dan Iancu} %\date{} %\keywords{robust optimization, multi-stage minimax, optimal policies, convex costs, dynamic, programming} %\maketitle \begin{center} {\Large Research Statement} \\[20pt] % {\Large Dan Iancu} \end{center} \thispagestyle{empty} % no numbering on first page either \vspace{-10pt} %\begin{abstract} %\end{abstract} \renewcommand*\thesubsection{} %\subsection{Motivation} \par %In today's increasingly globalized world, few would doubt the central role of operational decisions in the smooth functioning of a firm. Equally true is that, while important, these decisions are also becoming increasingly difficult, particularly when considering the complexities of the globalized business environment typical for today’s companies. %While responsible for many success stories, today's globalized business environment has also reemphasized the central role of operational decisions to the smooth functioning of a firm, and the fact that such decisions are becoming increasingly complex. %With the increased availability of data and inexpensive computing, recent years have witnessed a surge in the adoption of quantitative methods and tools aimed at automating as many decisions and processes as possible, across an ever-wider array of individual, business, and societal problems. %In parallel, the information technology sector has been responding to the ``analytics'' trend with numerous acquisitions and consolidations, betting on the fact that the vast amounts of data and increased computational power can generate advanced solutions for key business and societal problems. %While useful in simplifying routine processes, such solutions are nonetheless often limited in their scope, flexibility, and robustness. For instance, they often make unrealistic assumptions concerning the decision environment, using simplistic models of uncertainty, treating decisions in segregation, ignoring the existence of agency issues or alternative incentives, etc. Such simplifications are often a byproduct of pragmatism, sacrificing modeling realism for the sake of computational tractability. %Such tools are doubtlessly useful; but their underlying models and algorithms often entail pragmatic choices that render computational tractability at the cost of limiting the scope, flexibility, and robustness, by ignoring salient features on the information, incentives or broader environment surrounding the various decision makers. My research thus far has focused on providing new ways for understanding and computing optimal decisions and strategies in complex operational settings, in ways that not only support decision-making, but also render quantitative models more pertinent to real-world settings. My work has had two main thrusts: a more methodological one, aimed at developing new tools and algorithms for dynamic optimization under uncertainty and risk, and a more qualitative one, aimed at enhancing our understanding of the interplay between key operating decisions and important financial and contractual considerations surrounding a firm and its decision makers. Some of my early work on dynamic optimization under uncertainty has followed the \textbf{\textit{Robust Optimization (RO)}} paradigm, a methodology developed during the last 20 years to analyze and optimize the performance of complex systems under limited information.\footnote{The framework has a long history in various disciplines, including economics, engineering, and statistics, and is built around the premise that an exact probability distribution is often unavailable in a decision process due to, e.g., the lack of ample historical data or inherent non-stationarities. The standard approach is to assume that all unknown parameters belong to particular suitable uncertainty sets, and to seek feasible decisions that optimize the worst-case performance.} %Through recent developments in convex optimization, RO has emerged as an eminently tractable methodology for addressing real-world problems, with recent years witnessing an explosion of applications in management science.} %My primary focus has been on dynamic (i.e., multi-stage) problems, where certain decisions can be delayed until some uncertainty is resolved. In this context, I made the simple observation that %the well-known Bellman optimality principle of dynamic programming (DP) is no longer a {\it necessary} criterion for optimality in dynamic RO models \cite{DelIan}. Put differently, {\it multiple} solutions or policies typically exist in RO that achieve the same worst-case performance, but only some of which also satisfy the well-known Bellman optimality principle of dynamic programming (DP) \cite{DelIan}. This carries several important implications, which I discuss and explore in my research. On the one hand, I show how this degeneracy can help: since multiple optimal policies exist, some may also have a ``simple'' and intuitive structure, and be obtained by solving ``simple'' problems, without resorting to the potentially intractable DP recursions. To argue their existence though, one cannot appeal to usual DP arguments, and new techniques must be developed. In this spirit, in \cite{BerIanParMOR,IanShaSvi} I investigate several classical models for inventory management under limited demand information, for which I show that a new class of replenishment policies, which depend affinely on the observed demands, is actually optimal. From a \emph{theoretical} perspective, the result is the first of its kind, and the proof techniques are entirely novel, combining continuous optimization with lattice programming and discrete optimization to construct the optimal strategy. The result is also \emph{practically} relevant: such policies retain an intuitive interpretation (new orders are assigned to partially meet all backlogged demands), and optimal policies are easy to find, by solving a single linear optimization problem of small size. Interestingly, such ``simple'' policies remain optimal even if the cost structure becomes ``complex'' (e.g., arbitrary convex ordering costs), when the DP recursions become intractable or prescribe very complex policies. The approach can thus be adapted to more complex capacity planning problems under uncertainty, such as the design of flexible retailer-supplier contracts with pre-commitments and penalties \cite{IanShaSvi}. %\footnote{In follow-up work in [6], I also examine more complex, multi-tier supply chains, where I show how asymptotically optimal replenishment policies can be computed by solving tractable optimization problems, using techniques from polynomial and semidefinite optimization. The approach performs very well in simulation studies, and places a small burden on the decision maker, allowing a direct control of the trade-off between solution quality and computational complexity through one parameter.} Similarly, in follow-up work in [6], I show how a class of polynomial policies is provably optimal for more complex, multi-tier supply chains, and how such policies can be computed by solving a hierarchy of tractable, convex optimization problems. % At the heart of these results lies a simple -- but critical -- observation that my work highlighted. Namely, that the well-known Bellman optimality principle of dynamic optimization is a sufficient but {\it not} a necessary criterion for optimality in dynamic RO models, where the exact distribution is unknown % In this context, my work made a simple -- but somewhat critical -- observation that the well-known Bellman optimality principle, which is a necessary and sufficient governs optimal solutions to a dynamic problem under ample information (e.g., when exact distributions are known) %At the heart of these results lies a key observation concerning a critical distinction between decision making under severe uncertainty (i.e., RO) and decision making under ample information (i.e., exact distributions or deterministic setting). Namely, the former setting generally admits \emph{multiple} optimal policies, %\footnote{Intuitively, the Bellman optimality principle may be an unnecessarily stringent criterion for the former class of models, while it is strictly required for the latter. As such, robust decision models may admit structurally ``simpler'' optimal policies, and new methodology can be developed to find such policies, going beyond the direct use of DP.} with simpler structure than the latter. On the other hand, I also show how this degeneracy may cause serious inefficiencies and suboptimal performance when non-worst-case outcomes materialize. To this end, in [2], I introduce and formalize the concept of Pareto efficiency in RO. %The concept mimics the corresponding one in economics, engineering and multiobjective optimization. I show that decisions generated using classical RO methods need not have this property, and that particular algorithms/software used in RO problems are guaranteed to produce inefficient outcomes. I also provide simple procedures for testing whether a given solution is Pareto-efficient, and for generating Pareto-efficient solutions. This methodology is a strict improvement of the RO framework, incurring no extra computational cost, and significantly reducing conservativeness. %The ideas and arguments in the paper are completely novel, and the research has received the first prize in the INFORMS JFIG paper competition. Other related research on dynamic optimization under risk on which I do not elaborate for space reasons includes \cite{NegBimBraIan}---which develops adaptive treatment policies for chronic illnesses based on dynamically monitoring a new patient's state in order to reduce the substantial uncertainty related to her response to treatment---as well as [3]---which proposes ways to approximate a decision maker's static risk preferences (which may be easier to state) through axiomatically justified dynamic risk measures. Some of my active research continues to pursue such ideas on both a theoretical and an applied front. In \cite{IanWei}, I aim to generalize some of the insights in \cite{IanShaSvi} to a broader class of operational problems, by adapting ideas from the theory of discrete convexity. In \cite{BidIanSwi}, I explore the design of robust sourcing strategies in the presence of disruption risk and very limited information about the structure of a firm's deeper supply tiers (a situation that is common in today's globalized supply chains). Lastly, in \cite{IanTriYoo}, I consider the problem of determining a policy for monitoring and possibly altering the course of several processes with future unknown values, which govern a decision maker's payoffs; the problem has direct applications in collateralized lending (where a borrower pledges assets with uncertain values and the lender dynamically monitors the collateral value to assess whether it continues to cover the outstanding loan), but also in the design of medical treatments (where physicians determine the frequency of office visits or testing procedures....) Understanding how a firm's procurement and pricing decisions are influenced by the broader ecosystem surrounding the firm, including its suppliers and understanding how a firm's material and financial flows are intertwined %This underscores the fact that simplicity and robustness often go hand in hand: simple policies---such as those calculated under limited information---tend to be more robust than highly tailored ones---that rely on substantially more information. This concept, which is akin to the well-known concept of over-fitting is statics, %\textbf{\textit{Portfolio Management}}. Since the seminal work of Markowitz, multiple facets and extensions of this problem have been studied in the literature, and several software solutions now exist to guide trading decisions. An issue typically overlooked in most prior (implemented) models is the fact that a manager is usually in charge of several client accounts. In [1], I show that, due to market impact costs, this setting differs in several subtle ways from the classical (single account) case, with a key distinction arising from the performance of each account being coupled with the trading strategies of other accounts. I propose a novel approach for jointly determining the trades and splitting the associated market impact costs across all the accounts. The framework is flexible, allowing the manager to balance conflicting objectives, such as maximizing (risk-adjusted) aggregate wealth and distributing the gains equitably; it is also tractable, requiring convex optimization problems to be solved; furthermore, in numerical studies, it significantly outperforms existing methods employed in the industry or discussed in the literature.\footnote{Other related research, on which I do not elaborate for space reasons, includes [9], which develops optimal personalized treatments for chronic illnesses such as multiple sclerosis (through a quantitative model that explicitly uses all the channels available for acquiring information), as well as [3], which seeks to approximate a manager's static risk preferences (that may be simpler to formulate) by dynamic risk measures (that are axiomatically justified and computationally tractable.} In more recent years, my interests in risk management and global supply chains have also lead me to explore several In more recent work, I have also started examining several new issues that I believe are critical when developing operationally implementable solutions. The first is the role of agency issues, which I have explored in two classical operational contexts (sourcing in multi-tier supply chains [10], and inventory replenishment and liquidation under debt [8]). The second is the interplay between traditional operational decisions and financial, accounting or marketing considerations, which I have looked at in the context of financial covenants in inventory-based lending [8], as well as in ongoing research on dynamic pricing under debt [12] and the management of loyalty programs [11]. I briefly elaborate on some of these below. \textbf{\textit{Disruption Risks in Supply Chains}}. Despite extensive research in supply chain management, most models examining the contractual agreements in a complex supply chain typically ignore either the agency issues or the multi-tier structure of the chain (e.g., assuming a unique tier, of direct suppliers to a manufacturer). In [10], I discuss optimal sourcing decisions in a multi-tier chain, where the manufacturer is unable to directly dictate the sourcing decisions of its immediate Tier~1 suppliers (from a set of risky Tier 2 suppliers), but can influence their strategy through the contracts offered. I find that the manufacturer's optimal strategy critically depends on the degree of overlap in the supply chain, with move overlap inducing less reliance on direct mitigation (procuring excess inventory or multisourcing in Tier~1), and more on indirect mitigation (paying larger prices, and inducing Tier 1 suppliers to mitigate risk). While the manufacturer always prefers less overlap, Tier 1 suppliers may prefer a different configuration; this results in a severe moral hazard problem, which cannot be mitigated through side payments, but can be alleviated through penalty contracts. I also show that, when the supply chain structure is unknown, sourcing according to a robust strategy can significantly reduce the manufacturer's profit loss, particularly at low margins. \textbf{\textit{Financial Covenants in Inventory-based Lending}}. Classical papers focusing on operational decisions such as inventory replenishment or liquidation typically ignore the financing terms through which the inventory was purchased. In [8], I study the interplay between financial covenants and the operational decisions of a retailer who obtains financing through a secured (inventory-based) lending contract. While it is widely held that covenants serve to alleviate risk and protect lenders, the ways in which a retailer adapts his operations in response have not been studied. I show that retailers under debt manage inventory through riskier policies, involving surprising (non-threshold) behavior, and characterize the market conditions (involving demand distribution, growth potential, profit margin, inventory depreciation rate) under which covenants emerge as strictly necessary terms in lending agreements. I show that these conditions are routinely met in practice, and provide operational insights for the optimal design of covenants, arguing that additional operational flexibility can impact their effectiveness in a surprising, non-monotonic way. {\bf Nikos Ger} We study the inefficiencies stemming from a firm's operating flexibility under debt. We find that flexibility in replenishing or liquidating inventory, by providing risk shifting incentives, could lead to borrowing costs that erase more than a third of the firm’s value. In this context, we examine the e↵ectiveness of practical and widely used covenants in restoring firm value by limiting such risk shifting behavior. We find that simple financial covenants can fully restore value for a firm that possesses a mid-season inventory liquidation option. In the presence of added flexibility in replenishing or partially liquidating inventory, financial covenants fail, but simple borrowing base covenants successfully restore firm value. Explicitly characterizing optimal covenant tightness for all these cases, we find that better market conditions, such as lower inventory depreciation rate, higher gross margins or increased product demand, are typically associated with tighter covenants. Our results suggest that inventory-heavy firms can reap the full benefits of additional operating flexibility, irrespective of their leverage, by entering simple debt contracts of the type commonly employed in practice. For such contracts to be e↵ective, however, firms with enhanced flexibility and/or operating in better markets must also be willing to abide by more and/or tighter covenants. {\bf Omar} Firms often finance their inventory through debt and subsequently sell it to generate profits and service the debt. Pricing of products is consequently driven by both inventory and debt servicing considerations. In the present paper, we analyze how debt distorts dynamic pricing decisions and reduces generated sales revenues. We show that debt induces sellers to always price higher than the revenue-maximizing price. Furthermore, we establish that sellers under debt will always discount products at a lower pace than the revenue-maximizing one, and that the distortions and losses compound over time, leading to some form of performance spiral down. We quantify the extent to which such revenue losses can be mitigated by practical debt contract terms, which emerge as natural remedies from our analysis, and find debt amortization or financial covenants to be the most effective, followed by debt relief and early repayment options. {\bf Robert} We study sourcing in a supply chain with three levels: a manufacturer, Tier 1 suppliers, and Tier 2 suppliers prone to disruption from, e.g., natural disasters like earthquakes or floods. The manufacturer may not directly dictate which Tier 2 suppliers are used, but may influence the sourcing decisions of Tier 1 suppliers via contract parameters. The manufacturer’s optimal strategy depends critically on the degree of overlap in the supply chain: if Tier 1 suppliers share Tier 2 suppliers, resulting in a “diamond shaped” supply chain, the manufacturer relies less on direct mitigation (procuring excess inventory and multisourcing in Tier 1) and more on indirect mitigation (inducing Tier 1 suppliers to mitigate disruption risk). We also show that while the manufacturer always prefers less overlap, Tier 1 suppliers may prefer a more overlapped supply chain, and hence may strategically choose to form a diamond shaped supply chain. This preference conflict worsens as the manufacturer’s profit margin increases, as disruptions become more severe, and as unreliable Tier 2 suppliers become more heterogeneous in their probability of disruption; however, penalty contracts—in which the manufacturer penalizes Tier 1 suppliers for a failure to deliver ordered units—alleviate this coordination problem. {\bf Joann} In complex supply chains, the benefits and costs of technological innovations do not always accrue equitably to all parties; their adoption may thus critically depend on sourcing relationships and incentives. In a setting with uncertain and endogenous process yield, we study the potential of two features—contract design and sourcing channel—to create mutual benefit in decentralized value chains, where suppliers bear the costs of new technologies while benefits accrue primarily to buyers. Our focus is on agricultural value chains, where parties may transact through a channel that blends farmers’ produce (‘commodity-based channel’) or that allows a one-on-one interaction between farmer and processor (‘direct-sourcing channel’). Our study provides insights to companies seeking to incorporate responsible sourcing strategies while also creating economic value—a concept called ‘creating shared value.’ We identify that the technology’s ‘cost elasticity’ drives whether switching sourcing channel, changing contract structure, or adopting an integrated change is necessary to create shared value. This highlights that value chain innovations need to be properly designed— and sometimes combined—to achieve sustainable implementation. We also find that certain simple contracts with a linear or bonus structure are optimal, while other intuitive contracts could be detrimental. Using a dataset of farms in Patagonia, Argentina, we estimate that the proposed mechanism could increase average supply chain profit by 6.9\% while realizing positive environmental benefits. In my active research, I am following up on several of these problems, with the dual goal of (a) developing a better understanding of the incentive mechanisms at play in complex operational settings, and (b) developing tractable and implementable decision support tools. To give some concrete examples, in [12], [15] and [16], I am currently looking at the design of optimal monitoring policies, leveraging the use of IT systems to dynamically learn the operational status and performance of a borrower. In [11] and [17], I am looking at how the long-term design and management of loyalty and reward programs and of contracts for ad display is influenced by marketing, operational, but also financial and accounting considerations. Finally, in [13], I study the impact of climate risk on the design of long-term contracts, including financial subsidies for the adoption of sustainable practices in coffee and wool supply chains, and in [14] I look at policies for disaster relief in the Philippines, in the face of severe uncertainty (e.g., typhoon paths), and issues of equity in the availability of supplies. In addition to these problems, I also intend to pursue several methodological questions, related primarily to the development of specialized, computationally tractable, and data-driven tools for the design of robust long-term contractual agreements.\\[-10pt] \begin{center} {\Large Teaching Statement} \end{center} \vspace{-5pt} My primary teaching responsibilities have involved the MBA classes on ``Optimization and Simulation Modeling'' (OIT 245/247)%\footnote{OIT 245/247 are part of the foundational requirement for our first-year MBA students. Both courses teach basic skills for developing structured, quantitative models suitable for business decision making. Topics include basic optimization (linear, discrete, non-linear), as well as Monte-Carlo simulation and sensitivity analysis, all taught in a spreadsheet environment (currently, Microsoft Excel). Traditionally, both courses were offered during the first half of the winter quarter, in a very compact format (9 class sessions, of 145 minutes each), and consisted primarily of lecturing, coupled with some in-class exercises.} , taught jointly with my colleague, Kostas Bimpikis. In addition, I have also developed a new PhD class on inventory management (OIT 624), offered in alternate years. %I briefly discuss both of these below. During the past three years, Kostas and I have considerably reshaped the content, as well as the way in which OIT 245/247 are taught. As part of this effort, I developed six mini-cases that we currently use to illustrate applications of linear (discrete) optimization, and, in conjunction with Kostas, gradually proceeded to eliminate or substantially rewrite most of the other mini-cases, with the goal of showcasing more modern application areas of optimization and simulation. %\footnote{In particular, I developed mini-cases on: online advertising (``Marine Weekly'', 2012, reviewed in 2014), restaurant operations (``Impromptu'', 2012), optimizing family financials (``Family Financial Plan'', 2014), determining optimal diets (``Optimal Diet Problem'', 2014), rebalancing portfolios under transaction costs (``Portfolio Management'', 2014), engaging in store expansion and dual-channel operations (``Whole Wallet'', 2014), and helped write a mini-case on optimal sourcing and pricing for a coffee roasting business (``Starbucks Minicase'', 2013-2014).} During 2013, we also started exploring the use of educational technology, primarily with the goal of gradually ``flipping'' the classroom model. Initially, we developed videos that students could watch, either as preparation or to revisit concepts taught in class. Given the success of the videos, coupled with feedback from the students, I decided to explore an entirely different class format during winter 2014, moving away from a tiered classroom to a lab environment (the RAIL classroom, B312). Consequently, I also shifted the in-class focus away from lecturing, and more towards hands-on exercises, conducted in teams of two, under the supervision of the teaching staff. The class preparation was also reduced (students would watch videos or read short hand-outs, and turn in weekly assignments), and the new format allowed me to ensure that weaker students are able to follow the material, while keeping more advanced students engaged. %\footnote{All teams would be required to fill out a shared Google sheet, allowing me to track real-time performance and help teams that were falling behind, while also reserving 20-30 minutes at the end of class to discuss the solution and the ``general lessons''.} The new format seemed to work quite well, and we are currently implementing it across all sections of OIT 245/247. In the near future, I intend to continue experimenting with educational technology and new teaching paradigms (e.g., peer evaluations, a quantitative assignment of team composition, depending on students' backgrounds or progress, etc.), with the goal of making learning more fun and engaging, but also more effective. In terms of new material, I believe there is considerable scope for developing more classes on quantitative modeling. Over the past 2-3 iterations of OIT 245/247, I have been consistently faced with questions concerning follow-up options, only to point students to courses outside the GSB, and not entirely suited to their needs. Of key note is an advanced elective, combining optimization and simulation modeling with ideas from the Data \& Decisions classes (OIT 265/267). This new class could cover some new methodological tools (e.g., simple optimization under uncertainty, which we currently cannot teach in OIT 245/247 due to the timing of OIT 265/267), while devoting enough class time for projects around the use of data-driven analytical tools for real-world problems. Such a class would align our curriculum well with the growing focus (in industry and academia) on business analytics, while also addressing a gap in the current offerings. %Another potential offering in the MBA curriculum would be an advanced elective, focusing on the interplay between operational decisions and financial/accounting considerations. HBS currently offers such a class, and I believe this could be well tied with my own research agenda. In terms of PhD teaching, I intend to keep offering the class on inventory management, but in a slightly changed format, devoting half the class to traditional models and theory, and half to new directions for research. I would also like to offer a new PhD class on optimization methods, better suited to students in a business school. Topics would include fundamentals (linear and nonlinear optimization, duality concepts, KKT, combinatorial optimization, etc.), but also new topics (such as conic, semidefinite, polynomial or robust optimization), covering theoretical, as well as computational and implementation issues. \newpage \begin{thebibliography}{99} \small \bibitem{NegBimBraIan} D. Negoescu, K. Bimpikis, M. Brandeau, D. A. Iancu, ``Dynamic Learning of Patient Response Rates: An Application to Treating Chronic Diseases,'' {\it Forthcoming in} \textbf{Management Science}. \bibitem{AngIanSw} E. Ang, D. A. Iancu, and R. Swinney, ``Disruption Risk and Optimal Sourcing in Multi-tier Supply Networks,'' {\it Forthcoming in} \textbf{Management Science}. \vspace{-5pt} \bibitem{IanTrichTsou} D. A. Iancu, N. Trichakis, and G. Tsoukalas, ``Is Operating Flexibility Harmful Under Debt?,'' {\it Forthcoming in} \textbf{Management Science}. \bibitem{DelIan} E. Delage and D. A. Iancu, ``Robust Multi-stage Decision Making,'' \textbf{TutORials in Operations Research}, pp. 20-46, 2015. \bibitem{IanTrich} ``Fairness and Efficiency in Multiportfolio Optimization,'' D. A. Iancu and N. Trichakis, \emph{Forthcoming in \textbf{Operations Research}}. \bibitem{IanTrich} ``Pareto Efficiency in Robust Optimization,'' D. A. Iancu and N. Trichakis, \emph{\textbf{Management Science}, vol. 60, No. 1, pp. 130-147, 2014}. \bibitem{IanPetSub} ``Tight Approximations of Dynamic Risk Measures,'' D. A. Iancu, M. Petrik, and D. Subramanian, \emph{Forthcoming in \textbf{Mathematics of Operations Research}}. \bibitem{IanShaSvi} ``Supermodularity and Affine Policies in Dynamic Robust Optimization,'' D. A. Iancu, M. Sharma, and M. Sviridenko, \emph{\textbf{Operations Research}, vol. 61, No. 4, pp. 941-956, 2013}. %\bibitem ``A General Purpose Local Search Algorithm for Binary Optimization,'' D. Bertsimas, D. A. Iancu, and D. Katz-Rogozhnikov, \emph{\textbf{INFORMS Journal on Computing}, vol. 25, No. 2, pp. 208-221, 2013}. \bibitem{BerIanParIEEE} ``A Hierarchy of Near-Optimal Policies for Multi-stage Adaptive Optimization,'' D. Bertsimas, D. A. Iancu, and P. Parrilo. \emph{\textbf{IEEE Transactions on Automatic Control}, vol. 56, No. 12, pp. 2809-2824, 2011}. \bibitem{BerIanParMOR} ``Optimality of Affine Policies in Multi-stage Robust Optimization,'' D. Bertsimas, D. A. Iancu, and P. Parrilo. \emph{\textbf{Mathematics of Operations Research}, vol. 35, No. 2, pp. 363-394, 2010}. \bibitem ``Long-Term Management of Loyalty and Reward Programs,'' S-Y. Chun, D. A. Iancu, and N. Trichakis. \bibitem ``Dynamic Pricing Under Financial Covenants,'' O. Besbes, D. A. Iancu, and N. Trichakis. \bibitem ``Sustainability, Climate Risk, and Optimal Sourcing in Coffee Supply Chains,'' D. A. Iancu and Joann de Zegher. \bibitem ``Planning for Disaster Recovery and Relief in the Philippines,'' D. A. Iancu and J. Uichanco. \bibitem{IanTriYoo} Ian D. A. Iancu, N. Trichakis, and D.-Y. Yoon, ``Optimal Monitoring Policies Under Limited Information,'' {\it working paper}. \bibitem{IanWei} D. A. Iancu and Y. Wei, ``On the Role of Discrete Convexity in Robust Optimization.'' \bibitem{BidIanSwi} H. Bidkhori, D. A. Iancu, and R. Swinney, ``Managing Disruption Risk in Supply Chains with Limited Visibility.'' \bibitem{IanTur} ``Optimal Growth Policies in Online Advertising,'' D. A. Iancu, and J. Turner. \end{thebibliography} \small %\bibliographystyle{plainnat} %\bibliographystyle{amsplain} %\bibliography{optimal_markdown} %\bibliography{/Users/daniancu/Academic/MIT/Research/biblio} \end{document}