Matchmaking for Autonomous Agents in Electronic Marketplaces

摘要

Matchmaking is the process of mediating demand and supply based on profile information. Matchmaking plays a crucial role in agent-based electronic marketplaces: the problem to be solved is to find the most appropriate agents, products, or services for a task, negotiation, or market transaction. Most real-world problems require multidimensional matchmaking, I.e., the ability to combine various dimensions of decision-making to define an overall solution to a matchmaking problem, requiring the interplay of multiple matchmaking algorithms. In addition, in order to be applicable for real-world applications, the matchmaking component must be easily integrated into standard industrial marketplace platforms. The work described in this work aims at deploying agent-based matchmaking for industrial electronic business applications. The main contributions of this work are the following: (ⅰ) we provide a configurable framework called GRAPPA (Generic Request Architecture for Passive Provider Agents) which is designed to be adapted to electronic marketplace applications. Using GRAPPA, system designers can easily specify demand and supply profiles as XML objects; (ⅱ) within GRAPPA we provide an extensible library of matchmaking functions (building blocks) that can be used for rapid development of matchmaking solutions that include standard information retrieval algorithms.