Erick Delage

A unified framework for dynamic pari-mutuel information market design

Recently, coinciding with and perhaps driving the increased popularity of prediction markets, several novel pari-mutuel mechanisms have been developed such as the logarithmic market scoring rule (LMSR), the cost-function formulation of market makers, and the sequential convex parimutuel mechanism (SCPM). In this work, we present a unified convex optimization framework which connects these seemingly unrelated models for centrally organizing contingent claims markets. The existing mechanisms can be expressed in our unified framework using classic utility functions. We also show that this framework is equivalent to a convex risk minimization model for the market maker. This facilitates a better understanding of the risk attitudes adopted by various mechanisms. The utility framework also leads to easy implementation since we can now find the useful cost function of a market maker in polynomial time through the solution of a simple convex optimization problem. In addition to unifying and explaining the existing mechanisms, we use the generalized framework to derive necessary and sufficient conditions for many desirable properties of a prediction market mechanism such as proper scoring, truthful bidding (in a myopic sense), efficient computation, controllable risk-measure, and guarantees on the worst-case loss. As a result, we develop the first proper, truthful, risk controlled, loss-bounded (in number of states) mechanism; none of the previously proposed mechanisms possessed all these properties simultaneously. Thus, our work could provide an effective tool for designing new market mechanisms.

Percentile Optimization for Markov Decision Processes with Parameter Uncertainty

Markov decision processes are an effective tool in modeling decision making in uncertain dynamic environments. Because the parameters of these models typically are estimated from data or learned from experience, it is not surprising that the actual performance of a chosen strategy often differs significantly from the designers initial expectations due to unavoidable modeling ambiguity. In this paper, we present a set of percentile criteria that are conceptually natural and representative of the trade-off between optimistic and pessimistic views of the question. We study the use of these criteria under different forms of uncertainty for both the rewards and the transitions. Some forms are shown to be efficiently solvable and others highly intractable. In each case, we outline solution concepts that take parametric uncertainty into account in the process of decision making.

Automatic Single-Image 3d Reconstructions of Indoor Manhattan World Scenes

3d reconstruction from a single image is inherently an ambiguous problem. Yet when we look at a picture, we can often infer 3d information about the scene. Humans perform single-image 3d reconstructions by using a variety of single-image depth cues, for example, by recognizing objects and surfaces, and reasoning about how these surfaces are connected to each other. In this paper, we focus on the problem of automatic 3d reconstruction of indoor scenes, specifically ones (sometimes called “Manhattan worlds”) that consist mainly of orthogonal planes. We use a Markov random field (MRF) model to identify the different planes and edges in the scene, as well as their orientations. Then, an iterative optimization algorithm is applied to infer the most probable position of all the planes, and thereby obtain a 3d reconstruction. Our approach is fully automatic—given an input image, no human intervention is necessary to obtain an approximate 3d reconstruction.

Percentile optimization in uncertain Markov decision processes with application to efficient exploration

Markov decision processes are an effective tool in modeling decision-making in uncertain dynamic environments. Since the parameters of these models are typically estimated from data, learned from experience, or designed by hand, it is not surprising that the actual performance of a chosen strategy often significantly differs from the designers initial expectations due to unavoidable model uncertainty. In this paper, we present a percentile criterion that captures the trade-off between optimistic and pessimistic points of view on MDP with parameter uncertainty. We describe tractable methods that take parameter uncertainty into account in the process of decision making. Finally, we propose a cost-effective exploration strategy when it is possible to invest (money, time or computation efforts) in actions that will reduce the uncertainty in the parameters.

Decision Making Under Uncertainty When Preference Information Is Incomplete

We consider the problem of optimal decision making under uncertainty but assume that the decision makers utility function is not completely known. Instead, we consider all the utilities that meet some criteria, such as preferring certain lotteries over other lotteries and being risk averse, S-shaped, or prudent. These criteria extend the ones used in the first-and second-order stochastic dominance framework. We then give tractable formulations for such decision-making problems. We formulate them as robust utility maximization problems, as optimization problems with stochastic dominance constraints, and as robust certainty equivalent maximization problems. We use a portfolio allocation problem to illustrate our results. This paper was accepted by Dimitris Bertsimas, optimization.

Distributionally Robust Stochastic Knapsack Problem

This paper considers a distributionally robust version of a quadratic knapsack problem. In this model, a subsets of items is selected to maximizes the total profit while requiring that a set of knapsack constraints be satisfied with high probability. In contrast to the stochastic programming version of this problem, we assume that only part of the information on random data is known, i.e., the first and second moment of the random variables, their joint support, and possibly an independence assumption. As for the binary constraints, special interest is given to the corresponding semidefinite programming (SDP) relaxation. While in the case that the model only has a single knapsack constraint we present an SDP reformulation for this relaxation, the case of multiple knapsack constraints is more challenging. Instead, two tractable methods are presented for providing upper and lower bounds (with its associated conservative solution) on the SDP relaxation. An extensive computational study is given to illustrate...

Robust Partitioning for Stochastic Multivehicle Routing

The problem of coordinating a fleet of vehicles so that all demand points on a territory are serviced and the workload is most evenly distributed among the vehicles is a hard one. For this reason, it is often an effective strategy to first divide the service region and impose that each vehicle is only responsible for its own subregion. This heuristic also has the practical advantage that over time, drivers become more effective at serving their territory and customers. In this paper, we assume that client locations are unknown at the time of partitioning the territory and that each of them will be drawn identically and independently according to a distribution that is actually also unknown. In practice, it might be impossible to identify precisely the distribution if, for instance, information about the demand is limited to historical data. Our approach suggests partitioning the region with respect to the worst-case distribution that satisfies first- and second-order moments information. As a side product...

A Unified Framework for Dynamic Prediction Market Design

Recently, coinciding with and perhaps driving the increased popularity of prediction markets, several novel pari-mutuel mechanisms have been developed such as the logarithmic market-scoring rule (LMSR), the cost-function formulation of market makers, utility-based markets, and the sequential convex pari-mutuel mechanism (SCPM). In this work, we present a convex optimization framework that unifies these seemingly unrelated models for centrally organizing contingent claims markets. The existing mechanisms can be expressed in our unified framework by varying the choice of a concave value function. We show that this framework is equivalent to a convex risk minimization model for the market maker. This facilitates a better understanding of the risk attitudes adopted by various mechanisms. The unified framework also leads to easy implementation because we can now find the cost function of a market maker in polynomial time by solving a simple convex optimization problem. In addition to unifying and explaining the existing mechanisms, we use the generalized framework to derive necessary and sufficient conditions for many desirable properties of a prediction market mechanism such as proper scoring, truthful bidding (in a myopic sense), efficient computation, controllable risk measure, and guarantees on the worst-case loss. As a result, we develop the first proper, truthful, risk-controlled, loss-bounded (independent of the number of states) mechanism; none of the previously proposed mechanisms possessed all these properties simultaneously. Thus, our work provides an effective tool for designing new prediction market mechanisms. We also discuss possible applications of our framework to dynamic resource pricing and allocation in general trading markets.

Robust Optimization of Sums of Piecewise Linear Functions with Application to Inventory Problems

Robust optimization is a methodology that has gained a lot of attention in the recent years. This is mainly due to the simplicity of the modeling process and ease of resolution even for large scale models. Unfortunately, the second property is usually lost when the cost function that needs to be “robustified” is not concave (or linear) with respect to the perturbing parameters. In this paper we study robust optimization of sums of piecewise linear functions over polyhedral uncertainty set. Given that these problems are known to be intractable, we propose a new scheme for constructing conservative approximations based on the relaxation of an embedded mixed-integer linear program and relate this scheme to methods that are based on exploiting affine decision rules. Our new scheme gives rise to two tractable models that, respectively, take the shape of a linear program and a semidefinite program, with the latter having the potential to provide solutions of better quality than the former at the price of heavier computations. We present conditions under which our approximation models are exact. In particular, we are able to propose the first exact reformulations for a robust (and distributionally robust) multi-item newsvendor problem with budgeted uncertainty set and a reformulation for robust multiperiod inventory problems that is exact whether the uncertainty region reduces to a L 1 -norm ball or to a box. An extensive set of empirical results will illustrate the quality of the approximate solutions that are obtained using these two models on randomly generated instances of the latter problem.

The Value of Flexibility in Robust Location–Transportation Problems

This article studies a capacitated fixed-charge multiperiod location–transportation problem in which, while the location and capacity of each facility must be determined immediately, the determination of the final production and distribution of products can be delayed until actual orders are received in each period. In contexts where little is known about future demand, robust optimization, namely using a budgeted uncertainty set, becomes a natural method for identifying meaningful decisions. Unfortunately, it is well known that these types of multiperiod robust decision problems are computationally intractable. To overcome this difficulty, we propose a set of tractable conservative approximations for the problem that each exploit to a different extent the idea of reducing the flexibility of the delayed decisions. While all of these approximation models outperform previous approximation models that have been proposed for this problem, each also has the potential to reach a different level of compromise be...