General combinatorial auctions—auctions in which bidders place unrestricted bids for bundles of goods—are the subject of increasing study. Much of this work has focused on algorithms for finding an optimal or approximately optimal set of winning bids. Comparatively little attention has been paid to methodical evaluation and comparison of these algorithms. In particular, there has not been a systematic discussion of appropriate data sets that can serve as universally accepted and well motivated benchmarks. In this paper we present a suite of distribution families for generating realistic, economically motivated combinatorial bids in five broad real-world domains. We hope that this work will yield many comments, criticisms and extensions, bringing the community closer to a universal combinatorial auction test suite.

Standard auction mechanisms often break down in important e-commerce applications, where agents demand bundles of complementary resources, i.e. "I only want B if I also get A". This paper describes Bundle, an ascending-price auction that is guaranteed to compute optimal bundle allocations with agents that follow a best-response bidding strategy. The auction prices bundles directly and allows agents to place additive or exclusive-or bids over collections of bundles. Empirical results confirm that Bundle generates efficient allocations for hard resource allocation problems. Furthermore, we show that Bundle generates solutions without complete revelation (or computation) of agent preferences. Keywords Iterative auction, agent-mediated electronic commerce, resource allocation, bundling problem, price discrimination.

This paper presents eMediator, an electronic commerce server prototype that demonstrates ways in which algorithmic support and game-theoretic incentive engineering can jointly improve the efficiency of ecommerce. eAuctionHouse, the configurable auction server, includes a variety of generalized combinatorial auctions and exchanges, pricing schemes, bidding languages, mobile agents, and user support for choosing an auction type. We introduce two new logical bidding languages for combinatorial markets: the XOR bidding language and the OR-of-XORs bidding language. Unlike the traditional OR bidding language, these are fully expressive. They therefore enable the use of the Clarke-Groves pricing mechanism for motivating the bidders to bid truthfully. eAuctionHouse also supports supply/demand curve bidding. eCommitter, the leveled commitment contract optimizer, determines the optimal contract price and decommitting penalties for a variety of leveled commitment contracting mechanisms, taking into account that rational agents will decommit strategically in Nash equilibrium. It also determines the optimal decommitting strategies for any given leveled commitment contract. eExchangeHouse, the safe exchange planner, enables unenforced anonymous exchanges by dividing the exchange into chunks and sequencing those chunks to be delivered safely in alternation between the buyer and the seller.

Combinatorial auctions (CAs) have emerged as an important model in economics and show promise as a useful tool for tackling resource allocation in AI. Unfortunately, winner determination for CAs is NP-hard and recent algorithms have difficulty with problems involving goods and bids beyond the hundreds. We apply a new stochastic local search algorithm, Casanova, to this problem, and demonstrate that it finds high quality (even optimal) solutions much faster than recently proposed methods (up to several orders of magnitude), particularly for large problems. We also propose a logical language for naturally expressing combinatorial bids in which a single logical bid corresponds to a large (often exponential) number of explicit bids. We show that Casanovaperforms much better than systematic methods on such problems. 1

As traditional commerce moves on-line more business transactions will be mediated by software agents, and the ability of agent-mediated electronic marketplaces to efficiently allocate resources will be highly dependent on the complexity of the decision problems that agents face; determined in part by the structure of the marketplace, resource characteristics, and the nature of agents' local problems. We compare auction performance for agents that have hard local problems, and uncertain values for goods. Perhaps an agentmust solve a hard optimization problem to value a good, or interact with a busy and expensivehuman expert. Although auction design cannot simplify the valuation problem itself, we show that good auction design can simplify meta-deliberation -- providing incentives for the "right" agents to deliberate for the "right" amount of time. Empirical results for a particular cost-benefit model of deliberation show that an ascending-price auction will often support higher revenue and efficiency than other auction designs. The price provides agents with useful information about the value that other agents hold for the good.

This paper presents a decision theoretic framework that an autonomous agent can use to bid e#ectively across multiple, simultaneous auctions. Specifically, our framework enables an agent to make rational decisions about purchasing multiple goods from a series of auctions that operate di#erent protocols (we deal with the English, Dutch, First-Price Sealed Bid and Vickrey cases). The framework is then used to characterize the optimal decision that an agent should take. Finally, we develop a practical algorithm that provides a heuristic approximation to this ideal.

The Vickrey auction has been widely advocated for multiagent systems. First we review its limitations so as to guide practitioners in their decision of when to use that protocol. These limitations include lower revenue than alternative protocols, lying in non-private-value auctions, bidder collusion, a lying auctioneer, and undesirable revelation of sensitive information. We discuss the special characteristics of Internet auctions: third party auction servers, cryptography, and how proxy agents relate to the revelation principle and fail to promote truth-telling.

This article summarizes the bidding algorithms developed for the on-line Trading Agent Competition held in July, 2000 in Boston. At its heart, the article describes 12 of the 22 agent strategies in terms of (i) bidding strategy, (ii) allocation strategy, (iii) special approaches, and (iv) team motivations. The common and distinctive features of these agent strategies are highlighted. In addition, experimental results are presented that give some insights as to why the top-scoring agents' strategies were most eective

We present a computational study of an auction-based method for decentralized train scheduling. The method is well suited to the natural information and control structure of modern railroads. We assume separate network territories, with an autonomous dispatch agent responsible for the #ow of trains over each territory. Each train is represented by a self-interested agent that bids for the righttotravel across the network from its source to destination, submitting bids to multiple dispatch agents along its route as necessary. The bidding language allows trains to bid for the righttoenter and exit territories at particular times, and also to represent indi#erence over a range of times. Computational results on a simple network with straight-forward best-response bidding strategies demonstrate that the auction computes nearoptimal system-wide schedules. In addition, the method appears to have useful scaling properties, both with the number of trains and with the number of dispatchers, and generates less extremal solutions than those obtained using traditional centralized optimization techniques.

We develop a general framework in which real-time Dynamic Programming (DP) can be used to formulate agent bidding strategies in a broad class of auctions characterized by sequential bidding and continuous clearing. In this framework, states are represented primarily by an agent’s holdings, and transition probabilities are estimated from the market event history, along the lines of the “belief function ” approach of Gjerstad and Dickhaut [7]. We use the belief function, combined with a forecast of how it changes over time, as an approximate state-transition model in the DP formulation. The DP is then solved from scratch each time the agent has an opportunity to bid. The resulting algorithm optimizes cumulative long-term discounted profitability, whereas most previous strategies such as Gjerstad-Dickhaut (GD) merely optimize immediate profits. We test our algorithm in a simplified model of a Continuous Double Auction (CDA). Our results show that the DP-based approach reproduces the behavior of GD for small discount parameter γ, and is clearly superior for large values of γ close to 1. We suggest that this algorithm may offer the best performance of any published CDA bidding strategy. The framework our algorithm provides is extensible and can accommodate many market and research aspects.