A Multi-agent Based Negotiation Model With Fuzzy Logic Control

Along with the globalizing competition, it is hard for an enterprise to adapt market competition with limited manufacturing resource. The enterprises should cooperate with others to build up virtual enterprises, which form an agile supply chain to take the advantage of the partners. As virtual enterprise is composed and decomposed frequently, the important and difficulty work is how to reengineer agile supply chain with multi-agent negotiation rapidly and adaptively. Whereas, the negotiation is the first and crucial step to find and select suitable partners to establish agile supply chain.This paper focuses on the study of a multi-agent based negotiation model with fuzzy logic control. This model decompounds the cooperative partners of virtual enterprises as the roles of the agile supply through analyzing the business process of the agile supply chain; an agent stands for one role of the agile supply chain. We give the architecture of agent negotiation and the reasoning process model based multi-agent negotiation with fuzzy logic control after analyzing the process of multi-agent negotiation. An evaluating principle of the cooperative partners is given for this reasoning process model firstly, then, we analyze the factor, which affect the agent evaluating the cooperative partners with the principle. An evaluation index system is designed for multi-agent and multi-issue and multi-level negotiation, then, we evaluate the proposal in the agent negotiation with the index system fuzzily and synthetically and the result range from 0 to 1. This evaluation method is generic, furthermore, for the adaptability of agent, partiality of agent can be defined manually or automatically, the user can adjust the partiality with changing the weight to optimize negotiation results. The SAM principle is applied to determine the next state transfer to with the result. What's more, a counter-proposal is proposed with imitational means, then, the fuzzy automata of the agent for purchaser in the classic bi-negotiation is given based the above analysis. At last, we give an applied example of the negotiation model.This paper presents a new model for multi-agent, multi-issue and multi-level negotiation with fuzzy logic control, which is realistic, real-life, applied, generic, advantage-maximal, and Pareto-optimal. In a word, this paper establishes the academic base of designing an applied agent negotiation system. ...