基于Multi-agent的对策论协商模型  

A Game Theory Negotiation Model Based on Multi-Agent System

作　　者：张虹[1] 李晓[1] 邱玉辉[1] 

机构地区：[1]西南师范大学计算机科学与信息管理学院,重庆400715

出　　处：《计算机科学》2000年第12期86-88,共3页Computer Science

摘　　要：1 引言对策论的研究是以每个Agent的效用和赢利为基础的,每个Agent都希望能够在协商过程中求得Nash平衡的策略组合,使得那些基于理性假设的Agent均可获得一个稳定并且满意的解。当任何理性Agent达到Nash平衡点后,它们都不会试图离开,这是因为只有Nash平衡点,才能获得较大的赢利。可惜的是,Nash平衡不一定总是存在,或者可能存在多个,在这些情况下,如何寻找满意解。In traditional Game Theory,there do not exist or exist many Nash balance points. So it is very difficult to find a solution,which could make every agent to be satisfied. At a view of human's daily life ,this article tries to draw into a trusty third party to deal with some comparatively difficult game and condition,and draw into two communication behaviors: Asking Guarantee and Offering Compensation. That will make multi-agent find a compromise,stable ,comparatively good solution in negotiation,which would satisfy every agent.

关 键 词：对策论 协商模型 人工智能 MULTI-AGENT 

分 类 号：O225[理学—运筹学与控制论]