多 Agent 系统中的一种互助学习模型MSLM

A Model of Mutual Supervised Learning in Multiagent System MSLM

Ａｇｅｎｔ在与其他Ａｇｅｎｔ合作工作时，可以相互学习，以增强解决问题的能力．本文提出一种互助学习模型ＭＳＬＭ，在该模型中，每个Ａｇｅｎｔ均可作为其伙伴的教师，通过互帮互学，可以减少Ａｇｅｎｔ在解决相似问题时的重复协商时间，从而提高系统效率．

When cooperating with other agents,an agent can learn from each other.In this paper,we introduce a model of mutual supervised learning where every agent can act as the teacher of his partner. By mutual supervised learning,an agent can reduce the time of solving similar problems and improve the efficiency of the system.