Automated Negotiation Nased on Prioritised Fuzzy Constraint Satisfaction Problem

Automated negotiation is one of the key issues in e-commerce. For the purpose of better expressing Agent´s requirement and preference, easier dealing with issues inter-dependence and trade-off, this paper analyze fuzzy constraints´ significances to e-commerce negotiation detailedly, then proposed a general automated negotiation model based on prioritised fuzzy constraint satisfaction problem (PFCSP), including a formal model description and a new offer evaluation method based on fuzzy constraints´ satisfaction degree instead of utility theory. After that, solving PFCSP with genetic algorithm (GA) and developed a negotiation strategy based on constraint relaxation and GA, in which Agents can make concession by gradually relaxing his own fuzzy constraints, and generate the current best counter offer which decreased Agentspsila profit at the minimal extent. Finally, a negotiation example was given to describe the negotiation process and to demonstrate the availability of the model and the strategy.