Fuzzy Constraint-based Agents for Negotiation in e-Commerce

摘要

This paper presents a general problem-solving framework for the modeling of agent negotiation in electronic commerce via fuzzy constraint processing. We formulate agent negotiation as a distributed fuzzy constraint satisfaction problem (DFCSP). Fuzzy constraints, in this way, are used to represent the relationship among agents and define each agent's demands with human conceptualization. On the other hand, basing on fuzzy constraint-based problem-solving, we propose a concession strategy to relax demands and create new feasible proposals and also a trade-off strategy to evaluate the existing alternatives. This approach provides a systematic method to reach a satisfactory agreement without deleting any agent's demand arbitrarily or being annoyed by excessive feasible proposals. To demonstrate the effectiveness of our approach, an example application of e-Marketplace is provided.