MANAGERIAL ECONOMICS & DECISION SCIENCES
J. L. and Helen Kellogg Professor of Managerial Economics and Decision Sciences
We construct an ascending auction for heterogeneous objects by applying a primal-dual algorithm to a linear program that represents the efficient-allocation problem for this setting. The auction assigns personalized prices to bundles, and asks bidders to report their preferred bundles in each round. A bidder's prices are increased when he belongs to a "minimally undersupplied" set of bidders. This concept generalizes the notion of ``overdemanded" sets of objects introduced by Demange, Gale, and Sotomayor (1986) for the one-to-one assignment problem. Under a submodularity condition, the auction implements the Vickrey-Clarke-Groves outcome; we show that this type of condition is somewhat necessary to do so. When classifying the ascending-auction literature in terms of their underlying algorithms, our auction fills a gap in that literature. We relate our results to the recent work of Ausubel and Milgrom (2002).
The polyhedral structure of the K-median problem on a tree is examined. Even for very small connected graphs, we show that additional constraints are needed to describe the integer polytope. A complete description is given of those trees for which an optimal integer LP solution is guaranteed to exist. We present a new and simpler demonstration that an LP characterization of the 2-median problem is complete. Also, we provide a simpler proof of the value of a tight worst case bound for the LP relaxation. A new class of valid inequalities is identified. These inequalities describe a subclass of facets for the K-median problem on a general graph. Also, we provide polyhedral descriptions for several types of trees. As part of this work, we summarize most known results for the K-median problem on a tree.
A sequence of prices and demands are rationalizable if there exists a concave, continuous and monotone utility function such that the demands are the maximizers of the utility function over the budget set corresponding to the price. Afriat(1967) presented necessary and sufficient conditions for a finite sequence to be rationalizable. Varian(1982) and later Blundell et al. (2003, 2005) continued this line of work studying nonparametric methods to forecasts demand. Their methods do not implement any probabilistic model and therefore fall short of giving a general degree of confidence in the forecast. The present paper complements this line of research by introducing a statistical model and a measure of complexity through which we are able to study the learnability of classes of demand functions and derive a degree of confidence in the forecasts. In this paper we develop a framework to study the learnability of real vector valued demand functions through observations on prices and demand. Our results give lower and upper bounds on the sample complexity of PAC learnability and show that the sample complexity of learning a class of vector valued functions with finite fat shattering dimension increases by a linear factor of the dimension. We show that classes of income-Lipschitz demand functions with global bounds on the Lipschitz constant have finite fat shattering dimension.
We examine the mechanism design problem for a single buyer to procure purchase-options for a homogeneous good when that buyer is required to satisfy an unknown future demand. Suppliers have 2-dimensional types in the form of commitment costs and production costs. The efficient schedule of options depends on the distribution of demand. To implement an efficient outcome, we introduce a class of mechanisms which are essentially pivotal mechanisms (Vickrey--Clarke--Groves) with respect to the expected costs of the suppliers. We show that the computational task of running such mechanisms is not burdensome. Our discussion uses electricity markets as an example.
Each period an outcome (out of finitely many possibilities) is observed. For simplicity assume two possible outcomes, a and b. Each period, a forecaster announces the probability of a occurring next period based on the past. Consider an arbitrary subsequence of periods (e.g., odd periods, even periods, all periods in which b is observed, etc). Given an integer n, divide any such subsequence into associated sub-subsequences in which the forecast for a is between bounds. We compare the forecasts and the outcomes (realized next period) separately in each of these sub-subsequences. Given any countable partition of [0,1] and any countable collection of subsequences, we construct a forecasting scheme such that for all infinite strings of data, the long run average forecast for a matches the long run frequency of realized a's.
We characterize the class of Arrovian Social Welfare Functions (ASWFs) as integer solutions to a collection of linear inequalities. Many of the classical possibility, impossibility, and characterization results can be derived in a simple and unified way from this integer program. Among the new results we derive is a characterization of preference domains that admit a nondictatorial, neutral ASWF. We also give a polyhedral characterization of all ASWFs on single-peaked domains.
We study trading models when the distribution of signals such as costs or values is not known to traders or the mechanism designer when the profit-maximizing trading procedure is designed. We present adaptive mechanisms that simultaneously elicit this information (market research) while maintaining incentive compatibility and maximizing profits when the set of traders is large (market design). First, we study a monopoly pricing model where neither the seller nor the buyers know the distribution of values. Second, we study a model with a broker intermediating trade between a large number of buyers and sellers with private information about their valuations and costs. We show that when the set of traders becomes large our adaptive mechanisms achieve the same expected profits for the monopolist and the broker as when the distribution of signals is common knowledge.
We consider rules that choose a location on a graph (e.g. a road network) based on agents' single-peaked preferences. First, we characterize the class of strategy-proof, onto rules when the graph is a tree. Such a rule is based on a collection of generalized median voter rules (Moulin, 1980) satisfying a consistency condition. Second, we characterize such rules for graphs containing cycles. We show that while such a rule is not necessarily dictatorial, the existence of a cycle grants some agent an amount of decisive power, unlike the case of trees. Rules for this case can be described in terms of a subclass of such rules for trees.
A quitting game is a sequential game where each player has two actions: to continue or to quit. The game continues as long as all players decide to continue. The moment at least one player decides to quit, the game terminates. The terminal payoff depends on the subset of players who quit at the terminating stage. If the game continues forever, then the payoff for the players is some fixed-payoff vector. We prove that every quitting game admits a correlated uniform ε-equilibrium-a uniform ε-equilibrium in an extended game that includes a correlation device that sends one signal to each player before start of play
We introduce in this paper an exact nonlinear formulation of the multiway cut problem. By simple linearizations of this formulation, we derive several well-known and new formulations for the problem. We further establish a connection between the multiway cut and the maximum-weighted independent set problem. This leads to the study of several instances of the multiway cut problem through the theory of perfect graphs. We also introduce a new randomized rounding argument to study the sharpness of these formulations
In recent years, approximation algorithms based on randomized rounding of fractional optimal solutions have been applied to several classes of discrete optimization problems. In this paper, we describe a class of rounding methods that exploits the structure and geometry of the underlying problem to round fractional solution to 0–1 solution. This is achieved by introducing dependencies in the rounding process. We show that this technique can be used to establish the integrality of several classical polyhedra (min cut, uncapacitated lot-sizing, Boolean optimization, k-median on cycle) and produces an improved approximation bound for the min-k-sat problem.
At each point in time a decision maker must make a decision. The payoff in a period from the decision made depends on the decision as well as on the state of the world that obtains at that time. The difficulty is that the decision must be made in advance of any knowledge, even probabilistic, about which state of the world will obtain. A range of problems from a variety of disciplines can be framed in this way. In this paper we survey the main results obtained, as well as some of their applications.
In this paper we describe four axioms that uniquely characterize the class of locations in tree networks that are obtained by minimizing an additively separable, nonnegative, nondecreasing, differentiable, and strictly convex function of distances. This result is analogous to results that have been obtained in the theory of bargaining, social choice, and fair resource allocation.
Can we forecast the probability of an arbitrary sequence of events happening so that the stated probability of an event happening is close to its empirical probability? We can view this prediction problem as a game played against Nature, where at the beginning of the game Nature picks a data sequence and the forecaster picks a forecasting algorithm. If the forecaster is not allowed to randomise, then Nature wins; there will always be data for which the forecaster does poorly. This paper shows that, if the forecaster can randomise, the forecaster wins in the sense that the forecasted probabilities and the empirical probabilities can be made arbitrarily close to each other.
In the last 25 years approximation algorithms for discrete optimization problems have been in the center of research in the fields of mathematical programming and computer science. Recent results from computer science have identified barriers to the degree of approximability of discrete optimization problems unless P = NP. As a result, as far as negative results are concerned a unifying picture is emerging. On the other hand, as far as particular approximation algorithms for different problems are concerned, the picture is not very clear. Different algorithms work for different problems and the insights gained from a successful analysis of a particular problem rarely transfer to another. Our goal in this paper is to present a framework for the approximation of a class of integer programming problems (covering problems) through generic heuristics all based on rounding (deterministic using primal and dual information or randomized but with nonlinear rounding functions) of the optimal solution of a linear programming (LP) relaxation. We apply these generic heuristics to obtain in a systematic way many known as well as new results for the set covering, facility location, general covering, network design and cut covering problems.
Suppose two players repeatedly meet each other to play a game where 1. each uses a learning rule with the property that it is a calibrated forecast of the other's plays, and 2. each plays a myopic best response to this forecast distribution. Then, the limit points of the sequence of plays are correlated equilibria. In fact, for each correlated equilibrium there is some calibrated learning rule that the players can use which results in their playing this correlated equilibrium in the limit. Thus, the statistical concept of a calibration is strongly related to the game theoretic concept of correlated equilibrium.Journal of Economic LiteratureClassification Numbers: C72,D83,C44.
We consider the problem of minimizing the makespan when scheduling tasks on two uniform parallel machines, where one machine is q times as efficient on each task as is the other. We compute the maximum relative error of the LPT (largest processing time first) heuristic as an explicit function of q. In the special case that the two machines are identical (q = 1), our problem and heuristic reduce to the problem and heuristic analyzed by Graham (1969).
We describe axioms that uniquely characterize the absolute median and the squared median (the point that minimizes the sum of squared distances) of a tree. A common feature of these characterizations is a Consistency axiom of the type first used by H.P. Young to characterize a voting procedure called Borda's rule. These results make it possible to argue in favor of, or against, a particular location in terms of principles satisfied or violated.
We consider the on-line version of the original m-machine scheduling problem: given m machines and n positive real jobs, schedule the n jobs on the m machines so as to minimize the makespan, the completion time of the last job. In the on-line version, as soon as job j arrives, it must be assigned immediately to one of the m machines. We present two main results. The first is a (2 − ε)-competitive deterministic algorithm for all m. The competitive ratio of all previous algorithms approaches 2 as m → ∞. Indeed, the problem of improving the competitive ratio for large m had been open since 1966, when the first algorithm for this problem appeared. The second result is an optimal randomized algorithm for the case m = 2. To the best of our knowledge, our 4/3-competitive algorithm is the first specifically randomized algorithm for the original, m-machine, on-line scheduling problem.
Our main contribution is an O(n log n) algorithm that determines with high probability a perfect matching in a random 2-out bipartite graph. We also show that this algorithm runs in O(n) expected time. This algorithm can be used as a subroutine in an O(nsup2/sup) heuristic for the assignment problem. When the weights in the assignment problem are independently and uniformly distributed in the interval [0, 1], we prove that the expected weight of the assignment returned by this heuristic is bounded above by tex-math$3+O(n^{-a})$/tex-math, for some positive constant a.
Under a variety of different random models of the maximal covering location problem, we show that the relative error of a randomly generated solution converges to zero in expectation as problem size grows. We prove similar results for the relative error between the optimal integer and fractional solutions to the problem. Suppose that randomly generated instances of this problem are used to test heuristics. One consequence of our results is that we should expect the mean relative error of a heuristic to be better than that of randomly generated solutions, if the heuristic is to be considered useful.
We propose a randomized strategy for selecting/combining forecasts that is better than the forecasts used to produce it in a sense made precise in this paper. Unlike traditional methods this approach requires that no assumptions be made about the distribution of the event being forecasted or the error distribution and stationarity of the constituent forecasts. The method is simple and easy to implement.
The set covering problem has many diverse applications to problems arising in crew scheduling, facility location and other business areas. Since the problem is known to be hard to solve optimally, a number of approximate (heuristic) approaches have been designed for it. These approaches (with one exception) divide into two main groups, greedy heuristics and dual saturation heuristics. We use the concept of a Pareto optimal dual solution to show that an arbitrary dual saturation heuristic has the same worst-case performance guarantee as the two best known heuristics of that type. Moreover, this poor performance level is always attainable by those two heuristics.
This article presents a model involving employers and two classes of workers, alike except for labels. Employers choose whom to hire and workers choose whether to invest in training. At one equilibrium, employers discriminate, which, the authors show, is Pareto inferior to another equilibrium where no discrimination occurs. On the basis of this observation, an argument for affirmative action is advanced.
In this note, an $O ( | V |k )$ algorithm is described for determining whether an interval graph on $| V |$ vertices has a bandwidth less than or equal to a given integer $k$. While the algorithm is not the first to resolve this problem, it does admit a shorter proof of its correctness than a previous algorithm of the same complexity due to Kratsch (Information and Computation, 74 (1987), pp. 140–158)
We generalize the notion of a Condorcet point of a network by allowing an individual's vote to be weighted by an integral power of his/her distance from a facility to be sited. We show that on a tree network this ‘generalized’ Condorcet point coincides with the solution of a certain nonlinear single facility location problem. This result may be viewed as a generalization of an earlier result due to Hansen and Thisse (1985). We also give a polynomial algorithm for solving this nonlinear location problem.
Let N = (V, A) be a directed, arc weighted network with node set V and arc set A. Associated with each arc e ε A is a non-negative distance d(e) and a non-negative cost c(e) of removing it from N. Let B be the total amount available to spend on removing arcs. The most vital arcs problem (MVAP) is the problem of finding a subset K of arcs such that ΣeεKc(e) B and whose removal from N results in the greatest increase in the shortest distance between two specified nodes s and t in V. We show this problem to be NP-hard. In addition, we show that a closely related problem, whose solution provides a lower bound on the optimal solution value of MVAP, is polynomially solvable.
In this paper we study the probabilistic behavior of the first fit increasing heuristics for the dual bin packing problem. In this problem we seek to maximize the number of items that can be packed into m unit capacity bins. We show that when the items to be packed are independent and identically distributed, then the relative error of first fit increasing tends to zero in probability as the number of items tends to infinity. We also consider an on-line variant of this problem and propose a simple heuristics whose relative error tends to zero in probability as the number of items tends to infinity under the assumption that the item sizes are independently and identically distributed according to a distribution with finite mean and with zero in its compact support.
In this paper a linear time heuristic for a class of cyclic staffing problems with breaks is described. A novel feature of the heuristic is that the quality of the heuristic solution increases as the requirements for each period become more uniform. As a by-product of this work, a new class of polynomially solvable set-covering problems is identified.
In this paper we study the probabilistic behavior of the farthest neighbor heuristic for finding the longest Hamiltonian tour in a graph. We assume the edge weights are independent random variables uniformly distributed in [0,1]. If F, is the length of the heuristic tour and L, the optimal tour then Fn/Ln 1 a.s. as n . We also show that Ln/n 1 a.s. as n .
Orienteering is a sport in which start and end points are specified along with other locations. These other locations have associated scores. Competitors seek to visit, in a fixed amount of time, a subset of these locations on the way from the start point to the end point in order to maximize the total score. An effective center-of-gravity heuristic is presented that outperforms heuristics from the literature.
We study the sequential second price auction of multiple units of a homogeneous commodity. It is well known that such auctions can have inefficient equilibria. For the case of two bidders, we show that the value of the allocation obtained in a the unique subgame perfect equilibrium is at least $1-e^{-1}$ of the value of the efficient allocation. Furthermore, we show that this bound is asymptotically tight.
This paper introduces a strengthening of the notion of a stable core and characterizes it in terms of Kikiuta and Shapley's extendability condition.
Consider selling bundles of indivisible goods to buyers with concave utilities that are additively separable in money and goods. We propose an ascending auction for the case when the seller is constrained to sell bundles whose elements form a basis of a matroid. It extends easily to polymatroids. Applications include scheduling (Demange, Gale, and Sotomayor, 1986), allocation of homogeneous goods (Ausubel, 2004), and spatially distributed markets (Babaioff, Nisan, and Pavlov, 2004).Our ascending auction induces buyers to bid truthfully, and returns the economically efficient basis. Unlike other ascending auctions for this environment, ours runs in pseudo-polynomial or polynomial time. Furthermore we prove the impossibility of an ascending auction for nonmatroidal independence set-systems. [Extended abstract published as Bikhchandani, Sushil, Sven de Vries, James Schummer, and Rakesh V. Vohra (2008). Ascending auctions for integral (poly)matroids with concave nondecreasing separable values. In SODA ’08: Proceedings of the nineteenth annual ACM-SIAM symposium on Discrete algorithms. Society for Industrial and Applied Mathematics, Philadelphia, PA, USA, pages 864–873.]
We examine the problem of allocating a resource repeatedly over time amongst a set of agents. The utility that each agent derives from the consumption of the item is private information to that agent and, prior to consumption, may be unknown even to that agent. The problem is motivated by keyword auctions, where the resource to be allocated is a slot on a search page. We describe a mechanism based on a sampling-based learning algorithm that under suitable assumptions is asymptotically individually rational, asymptotically Bayesian incentive compatible and asymptotically ex-ante efficient.
We consider an environment with a single divisible good and two bidders. The valuations of the bidders are private information but one bidder has a commonly known budget constraint. For this environment we derive the revenue maximizing subsidy free incentive compatible auction. We also examine the case when the budget constraint is private information but bidders must post a bond.
Despite impossibility results on general domains, there are some classes of situations in which there exist interesting dominant-strategy mechanisms. While some of these situations (and the resulting mechanisms) involve the transfer of money, we examine some that do not. Specifically, we analyze: problems where agents have single-peaked preferences over a 1-dimensional "public" policy space; problems where agents can trade/consume a single, indivisible "private" good; and problems where agents must match with each other.
The Vickrey sealed bid auction occupies a central place in auction theory because of its efficiency and incentive properties. Implementing the auction requires the auctioneer to solve n+1 optimization problems, where n is the number of bidders. In this paper we survey various environments (some old and some new) where the payments bidders make under the Vickrey auction correspond to dual variables in certain linear programs. Thus, in these environments, at most two optimization problems must be solved to determine the Vickrey outcome. Furthermore, primal-dual algorithms for some of these linear programs suggest ascending auctions that implement the Vickrey outcome.
The incentive compatibility constraints of mechanism design (with quasi-linear utilities) can be interpreted as being the dual inequalities of a shortest path problem. I believe this interpretation provides an intuitive and powerful lens through which to view the implications of incentive compatibility. This document provides an (incomplete) account of the main results in mechanism design from this point of view. They assume familiarity with mechanism design and basic linear programming and convex analysis. I welcome suggestions for additions, deletions and improvements to this material.
This course counts toward the following majors: Decision Sciences.
Provides frameworks for reasoning about decisions in uncertain environments. Case studies and experiments are used to motivate the importance of probabilistic reasoning to avoid the systematic biases that cloud managers' decision making. Formal probabilistic tools are introduced and their relevance to modern business issues is conveyed via cases, exercises, and class experiments. Some of the applications include: inventory management with uncertain demand, principal-agent models, herd behavior, selection bias, rare events, real options and risk. The course is self-contained, and should be of value to all students, including those with prior exposure to formal probability models.
Statistical Methods For Management Decisions (DECS-434-0)
This course counts toward the following majors: Decision Sciences.
This sequel to DECS-433 extends the statistical techniques learned in that course to allow for the exploration of relationships between variables. Topics include one- and two-population hypothesis testing, correlation, simple and multiple regression analysis, and qualitative variables. The course also covers applications of the material and a number of case studies. Extensive use of spreadsheet statistical analysis software is required.
Pricing Strategies (MECN-446-0)
This course counts toward the following majors: Analytical Consulting, Decision Sciences, Entrepreneurship & Innovation, Managerial Economics.
This course provides students with a comprehensive framework for formulating and implementing pricing strategies. Techniques that take account of the often imprecise and uncertain information to which management has access are useded to analyze the influence of costs, demand and competition on the pricing decision. Also discussed are research methods that can complement managerial judgment and the importance of maintaining consistency with other elements of the marketing mix. Special attention is devoted to the design of pricing schemes that segment the market, such as peak-load pricing, product bundling and nonlinear pricing. The course also studies vertical pricing problems (transfer pricing, pricing and distribution), legal constraints on pricing and industrial pricing (bidding). Actual pricing schemes in various industries and selected cases are used for illustrative and analytical purposes.
Prerequisites: MECN-430, MKTG-430.
PHONE: 847-491-5428
FAX: 847-467-1220
Jacobs Center Room 597