On-Line Dynamic Reinforcement Learning and Its Application in Multi-agent Based E-commerce

摘要：

With the rapid development of multi-agent systems (MAS),on-line automatic negotiation is often needed.But because of incomplete information agents have,the efficiency of on-1ine negotiation is rather low.To overcome the problem.on-line reinforcement learning algorithm is presented to learning the incomplete information of negotiation agent to enhance the emciency of negotiation.The algorithm is applied to on-line bilateral multi-issue negotiation in Multi-agent based electronic commerce.Three kinds of agents are used to compare with,which are no-learning agents(NA),static learning agents(SA) and dynamic learning agent (DA)in this paper.In static learning agent,the learning rate of Q-learning-s set to O.1unchangeable,so its called static Iearning.WIIiIe in dynamic learning proposed by this paper,the learning rate of Q-learning can change dynamicaHy.so.t’s called dynamic learning.Experiments show that it can help agents to negotiate more efficiently.

作者: LI Jian JING Bo

会议类型: 国际会议

会议名称: The International Conference Information Computing and Automation(2007国际信息计算与自动化会议)

会议地点: 成都

会议语种:英文

页码: 1359-1362

在线出版日期: 2007-12-19（万方平台首次上网日期，不代表论文的发表时间）