摘要
为了帮助协商Agent选择最优行动实现其最终目标,提出基于贝叶斯分类的增强学习协商策略。在协商过程中,协商Agent根据对手历史信息,利用贝叶斯分类确定对手类型,并及时动态地调整协商Agent对对手的信念。协商Agent通过不断修正对对手的信念,来加快协商解的收敛并获得更优的协商解。最后通过实验验证了策略的有效性和可用性。
To help negotiation Agent to select its best actions and reach its final goal,a reinforcement learning negotiation strategy based on Bayesian classification was proposed.In the middle of negotiation process,negotiation Agent makes the best use of the opponent＇s negotiation history to make a decision of the opponent＇s type based on Bayesian classification,dynamically adjust the negotiation Agent＇s belief of opponent in time,quicken the negotiation result convergence and reach the better negotiation result.Finally,the algorithm was proved to be effective and practical by experi-ment.
出处
《计算机科学》
CSCD
北大核心
2011年第9期227-229,247,共4页
Computer Science
基金
中央高校基本科研业务费科研专项项目(CDJRC10180012
CDJZR10180014)资助
关键词
贝叶斯分类
增强学习
协商策略
协商历史
Bayesian classification
Reinforcement learning
Negotiation strategy
Negotiation history
