多目标进化算法在主体自动协商中的应用

摘要

Based on the research of the negotiation strategy and communication protocol between agents in multi-agents system, the one to one negotiation model is analyzed. Response strategy and think strategy are combined in the proposed model in this paper, fully account the preference, maximize the utility value of each sides. Then use the multi-objection evolutionary to establish the pre-providing proposal set and the dominated set. Multi-agent auto-negotiation is realized. Experimental results verify the validity of this method.%研究多主体系统中主体之间的协商协议和通信协议,分析一对一协商模型,该模型采用响应型策略和慎思型策略进行协商.考虑协商双方的偏好和最大化己方的效用值,建立多目标进化算法,构造主体的预提供方案集及其支配子集,实现多主体自动协商.实验结果验证了该方法的有效性.