We study bargaining networks, discussed in a recent paper of Kleinberg and Tardos, from the perspective of cooperative game theory. In particular we examine three solution concepts, the nucleolus, the core center and the core median. All solution concepts define unique solutions, so they provide testable predictions. We define a new monotonicity property that is a natural axiom of any bargaining game solution, and we prove that all three of them satisfy this monotonicity property. This is actually in contrast to the conventional wisdom for general cooperative games that monotonicity and the core condition (which is a basic property that all three of them satisfy) are incompatible with each other. Our proofs are based on a primal-dual argument (for the nucleolus) and on the FKG inequality (for the core center and the core median). We further observe some qualitative differences between the solution concepts. In particular, there are cases where a strict version of our monotonicity property is a natural axiom, but only the core center and the core median satisfy it. On the other hand, the nucleolus is easy to compute, whereas computing the core center or the core median is #P-hard (yet it can be approximated in polynomial time).

Bargaining games on exchange networks have been studied by both economists and sociologists. A Balanced Out- come for such a game is an equilibrium concept that combines notions of stability and fairness. In a recent paper, Kleinberg and Tardos introduced balanced outcomes to the computer science community and provided a polynomial-time algorithm to compute the set of such outcomes. Their work left open a pertinent question: are there natural, local dynamics that converge quickly to a balanced outcome? In this paper, we provide a partial answer to this question by showing that simple edge- balancing dynamics converge to a balanced outcome whenever one exists.

We analyze the problem of designing a truthful pay-per-click auction where the click-through-rates (CTR) of the bidders are unknown to the auction. Such an auction faces the classic explore/exploit dilemma: while gathering information about the click through rates of advertisers, the mechanism may loose revenue; however, this gleaned information may prove valuable in the future for a more profitable allocation. In this sense, such mechanisms are prime candidates to be designed using multi-armed bandit techniques. However, a naive application of multi-armed bandit algorithms would not take into account the strategic considerations of the players --- players might manipulate their bids (which determine the auction's revenue) in a way as to maximize their own utility. Hence, we consider the natural restriction that the auction be truthful. The revenue that we could hope to achieve is the expected revenue of a Vickrey auction that knows the true CTRs, and we define the truthful regret to be the difference between the expected revenue of the auction and this Vickrey revenue. This work sharply characterizes what regret is achievable, under a truthful restriction. We show that this truthful restriction imposes statistical limits on the achievable regret --- the achievable regret is $\tilde{\Theta}(T^{2/3})$, while for traditional bandit algorithms (without the truthful restriction) the achievable regret is $\tilde{\Theta}(T^{1/2})$ (where $T$ is the number of rounds). We term the extra $T^{1/6}$ factor, the `price of truthfulness'.

We consider the problem of a search engine trying to assign a sequence of search keywords to a set of competing bidders, each with a daily spending limit. The goal is to maximize the revenue generated by these keyword sales, bearing in mind that, as some bidders may eventually exceed their budget, not all keywords should be sold to the highest bidder. We assume that the sequence of keywords (or equivalently, of bids) is revealed on-line. Our concern will be the competitive ratio for this problem versus the off-line optimum. We extend the current literature on this problem by considering the setting where the keywords arrive in a random order. In this setting we are able to achieve a competitive ratio of $1-\epsilon$ under some mild, but necessary, assumptions. In contrast, it is already known that when the keywords arrive in an adversarial order, the best competitive ratio is bounded away from 1. Our algorithm is motivated by PAC learning, and proceeds in two parts: a training phase, and an exploitation phase.

We study auctions for selling a limited supply of a single commodity in the case where the supply is known in advance and the case it is unknown and must be instead allocated in an online fashion. The latter variant was proposed by Mahdian and Saberi as a model of an important phenomena in auctions for selling Internet advertising: advertising impressions must be allocated as they arrive and the total quantity available is unknown in advance. We describe the Bayesian optimal mechanism for these variants and extend the random sampling auction of Goldberg et al. to address the prior-free case.

We consider markets in the classical Arrow-Debreu model. There are n agents and m goods. Each buyer has a concave utility function (of the bundle of goods he/she buys) and an initial bundle. At an ``equilibrium'' set of prices for goods, if each individual buyer separately exchanges the initial bundle for an optimal bundle at the set prices, the market clears, i.e., all goods are exactly consumed. Classical theorems guarantee the existence of equilibria, but computing them has been the subject of much recent research. In the related area of Multi-Agent Games, much attention has been paid to the complexity as well as algorithms. While most general problems are hard, polynomial time algorithms have been developed for restricted classes of games, when one assumes the number of strategies is constant.

For the Market Equilibrium problem, several important special cases of utility functions have been tackled. Here we begin a program for this problem similar to that for multi-agent games, where general utilities are considered. We begin by showing that if the utilities are separable piece-wise linear concave (PLC) functions, and the number of goods (or alternatively the number of buyers) is constant, then we can compute an exact equilibrium in polynomial time. Our technique for the constant number of goods is to decompose the space of price vectors into cells using certain hyperplanes, so that in each cell, each buyer's threshold marginal utility is known. Still, one needs to solve a linear optimization problem in each cell. We then show the main result - that for general (non-separable) PLC utilities, an exact equilibrium can be found in polynomial time provided the number of goods is constant. The starting point of the algorithm is a ``cell-decomposition'' of the space of price vectors using polynomial surfaces (instead of hyperplanes). We use results from computational algebraic geometry to bound the number of such cells. For solving the problem inside each cell, we introduce and use a novel LP-duality based method. We note that if the number of buyers and agents both can vary, the problem is PPAD hard even for the very special case of PLC utilities such as Leontief utilities and separable PLC utilities.

Determining the integrality gap of the bidirected cut relaxation for the metric Steiner tree problem, and exploiting it algorithmically, is a long-standing open problem. We use geometry to define an LP whose dual is equivalent to this relaxation. This opens up the possibility of using the primal-dual schema in a geometric setting for designing an algorithm for this problem. Using this approach, we obtain a 4/3 factor algorithm and integrality gap bound for the case of quasi-bipartite graphs; the previous best integrality gap upper bound being 3/2. We also obtain a factor $sqrt{2}$ strongly polynomial algorithm for this class of graphs. A key difficulty experienced by researchers in working with the bidirected cut relaxation was that any reasonable dual growth procedure produces extremely unwieldy dual solutions. A new algorithmic idea helps finesse this difficulty - that of reducing the cost of certain edges and constructing the dual in this altered instance - and this idea can be extracted into a new technique for running the primal-dual schema in the setting of approximation algorithms.

Arora, Rao and Vazirani showed that the standard semi-definite programming (SDP) relaxation of the sparsest cut problem with the triangle inequality constraints has an integrality gap of $O(\sqrt{\log n})$. They conjectured that the gap is bounded from above by a constant. In this paper, we disprove this conjecture by constructing an $\Omega(\log \log n)$ integrality gap instance. Khot and Vishnoi had earlier disproved the non-uniform version of this Conjecture. A simple ``stretching'' of the integrality gap instance for the sparsest cut problem serves as an $\Omega(\log \log n)$ integrality gap instance for the SDP relaxation of the Minimum Linear Arrangement problem. This SDP relaxation was considered in Charikar et. al. and Feige and Lee, where it was shown that its integrality gap is bounded from above by $O(\sqrt{\log n} \log \log n).$

The traditional model of market equilibrium supports impressive existence results, including the celebrated Arrow-Debreu Theorem. However, in this model, polynomial time algorithms for computing (or approximating) equilibria are known only for linear utility functions. We present a new, and natural, model of market equilibrium that not only admits existence and uniqueness results paralleling those for the traditional model but is also amenable to efficient algorithms.

Strategyproof cost-sharing mechanisms, lying in the core, that recover $1/\alpha$ fraction of the cost, are presented for the set cover and facility location games; $\alpha = O(\log n)$ for the former and 1.861 for the latter. Our mechanisms utilize approximation algorithms for these problems based on the method of dual-fitting.

We give the first polynomial time algorithm for exactly computing an equilibrium for the linear utilities case of the market model defined by Fisher. Our algorithm uses the primal-dual paradigm in the enhanced setting of KKT conditions and convex programs. We pinpoint the added difficulty raised by this setting and the manner in which our algorithm circumvents it.