摘要
提出一种具有自主学习能力的并发协商模型,通过使用增强学习方法的Q学习算法生成协商提议,使用相似度方法评价提议,使得Agent能够在半竞争、信息不完全和不确定以及存在最大协商时间的情况下,更为有效地完成多议题多Agent并发协商。
A concurrent negotiation model with automated learning capability was developed in trading environments. By using Q-learning algorithm to propose own proposal and similarity criteria to evaluate the opposing party＇s proposal, agents can participate in concurrent multi-issue negotiation in semi-competitive situations in which there exists information uncertainty and deadlines. This model enables the negotiation more effective.
出处
《计算机应用》
CSCD
北大核心
2006年第3期663-665,共3页
journal of Computer Applications
关键词
并发协商
自动协商
增强学习
Q学习
相似度方法
concurrent negotiation
automated negotiation
reinforcement leaming
Q-learning
similarity criteria
