We consider auction design in a setting with costly preference elicitation. We motivate the role of proxy agents, that are situated between bidders and the auction, and maintain partial information about agent preferences and compute equilibrium bidding strategies based on the available information. The proxy agents can also elicit additional preference information incrementally during an auction. We show that indirect mechanisms, such as proxied ascending-price auctions, can achieve better allocative efficiency with less preference elicitation than direct mechanisms, such as sealed-bid auctions.

Machine-readable specifications of auction mechanisms facilitate configurable implementation of computational markets, as well as standardization and formalization of the auction design space. We present an implemented rule-based scripting language for auctions, which provides constructs for specifying temporal control structure, while supporting orthogonal definition of mechanism policy parameters. Through a series of examples, we show how the language can capture much of the space of single-dimensional auctions, and can be extended to cover other novel designs.

Abstract. The Trading Agent Competition (TAC) is an international forum which promotes high-quality research regarding the trading agent problem. One of the TAC competitive scenarios is Supply Chain Management (SCM) where six agents compete by buying components, assembling PCs from these components and selling the manufactured PCs to customers. The idea of TAC SCM is the development of agent-based architectures that implement wideapplicable business strategies which efficiently manage cross-enterprise processes. In this paper, we analyze the TAC SCM environment and describe the main features of the CrocodileAgent, our entry in the TAC SCM Competition. 1

Abstract. Many problems studied in the multiagent systems community can be considered instances of an abstract multiagent resource allocation problem. In this problem, which is now better understood theoretically, the goal is to satisfy a criterion of global optimality (formulated in terms of a suitable notion of social welfare), given that the agents sharing a set of resources follow a local rationality criterion reflecting their individual preferences. In this paper, we first show that this simple decentralised resource allocation framework allows us to model a wide variety of applications. These applications thereby benefit from all the theoretical results concerning the framework. We then draw up a list of criteria which can guide the application designer working within the framework and illustrate the relevance of our approach by discussing several applications in view of this list of design criteria. 1

In previous work we proposed an approach for computing an agent’s preferences over different schedules of tasks, and for soliciting desirable bid combinations to cover the tasks. The proposed approach finds schedules that maximize the agent’s Expected Utility. The maximization problem is hard because the domain is piece-wise continuous, with the number of pieces and local maxima growing exponentially in the worst case scenario. For agents who are averse to taking risks, maximization algorithms tend to converge to degenerate maxima of no practical interest. In this paper we demonstrate three different maximization methods based on domain-specific heuristics. We also present a new stochastic maximization approach and benchmark it in two substantially different problem setups. 1.

This note describes a sample implementation of automated negotiations in an e-commerce modeling multi-agent system. A specific set of rules is used for enforcing negotiation mechanisms. Discussion of system design and implementation using JADE and JESS is provided. Finally, an experiment involving multiple English auctions performed in parallel is discussed.

Abstract Peer-to-peer markets going mobile spur spontaneity in trading considerably. Spontaneity, however, imposes severe informational requirements on the market participants. Informational requirements are twofold: Firstly, participants have to agree on a common vocabulary for that spontaneous market. Secondly, they need precise information about how the trading process is organized. Due to the lack of a central market operator these common understandings must be determined by the market participants themselves. Prior to any market process, these terms and regulations must be distributed by the market participant that initiates the market process. This raises the question concerning the used ontology. Standards describing (business) processes are available in general, but are currently not suitable for ephemeral markets. Electronic markets are extremely context sensitive, making the establishment of common understandings crucial as well as difficult. This paper uses structural similarities of markets and creates a Minimal Market Model. Since this model is for all conceivable markets constituent, it can be used as the core component. As any market is founded on this minimal model, it can be systematically extended to capture the peculiarities of each particular market. The derivation of the Minimal Market Model is founded on a solid ground of economic theory and refined such that it can be expressed in a formal way. 1

Abstract. With the spread of e-commerce on a global scale, the development of truly open semantic descriptions of negotiation mechanisms for agent systems generated a lot of interest in the research community. This paper proposes the use of the REWERSE rule-markup language R2ML for semantic modeling of negotiation mechanisms to enable agents to engage in more flexible and open negotiations. Rules are developed on top of an ontology of negotiation concepts and define a lingua franca for all software agents participating in negotiation. 1

Abstract. Mediation between buyers (service requester’s agents) and sellers (service provider’s agents) is one of the most difficult problems in real electronic markets. In this paper, we propose an economic approach to solving this problem combined with AI (Artificial Intelligence) concepts. Firstly, we enable provider agents to dynamically and autonomously advertise semantic descriptions of available services by proposing a new auction model based on Pay-Per-Click advertising auctions. We call it the Semantic Pay-Per-Click Agent (SPPCA) auction. Requester agents then use two-level filtration of the advertised services to efficiently discover eligible services. In the first level of filtration, a semantic-based mechanism for matchmaking between services requested by buyers and those advertised by sellers is applied. Services which pass the first level of filtration are then considered on the second level. Here information regarding the actual performance of service providers is considered in conjunction with the prices bid by service provider’s agents in the SPPCA auction. A final set of advertised services is then chosen and proposed to the buyer agent as an answer to its request. Keywords: Intelligent software agents, Web services, the semantic Web, OWL-S, Pay-Per-Click (PPC) auctions, digital economy, electronic commerce

Abstract not found