基于贝叶斯决策的虚拟企业多智能体协商研究

虚拟企业是不同于传统企业的敏捷化组织形式,基于多智能体的协商被用于处理虚拟企业中的协作和冲突消解等问题.文章引入贝叶斯决策方法,建立自学习协商模型.根据协商中的更新信息,实时更新智能体信念,实现多智能体协商过程的自学习功能.从初步试验结果可知,模型提高了协商效率,增加了协商智能体的效用.

Bayesian-decision-based Multi-agent Negotiations in Virtual Enterprises

As virtual enterprises differ from oridinary ones in their promptness of organization, the multi-agent negotiation has been adopted to mediate and resolve conflicts therein. In this paper, we propose a model of multi-agent self-learning and negotiation based on the Bayesian decision method, which can fulfill the real-time update of the belief of the agent in accordance with messages from the negotiation, thereby facilitating the self-learning in the multi-agent negotiation. Initial experimental results indicate that the model improves the efficiency of negotiations and increases the utility.