Incremental Learning and Its Application in Multi-agent Based E-commerce
With the rapid development of MAS,the on-line negotiation is needed urgently.But because of the incomplete information,the negotiation is rather inefficient.In this paper,an on-line dynamic incremental learnng algorithm is presented to overcome the problem,which can learn the agent’s reservation price by Q-learning algorithm.During the negotiation process.the agent which uses the algorithm is in the dynamic environment,and need not to save the negotiation information.so it is on-line.The agent learns the incomplete information by the feedback information through negotiation.In the experiment,the two kinds of agent are used to compare,one is the agent with no learning,and the other is the agent with the learning algorithm in this paper.The experiment results show that the agents with the learning algorithm in this paper can learn the incomplete information of its opponent and so they can negotiate more efficiently.
LI Jian JING Bo
国际会议
The International Conference Information Computing and Automation(2007国际信息计算与自动化会议)
成都
英文
867-869
2007-12-19（万方平台首次上网日期，不代表论文的发表时间）