【摘要】 Agent具有自私的特性.为了在多agent系统学习时使学习效果体现最大的群体利益,提出了一种改进的多agent学习方法.该学习方法改进了基于基组织共识学习的学习方法,在多agent学习时考虑了主体的可信度问题,从而使最后的学习结果更具真实性.更多还原

【Abstract】 Agent is selfish.In order to get maximal system benefit through MAS learning,an improved approach of learning in Multi-agent system is proposed.In this approach,the learning method which based on organization-structure by considering the credit of the agent,and gets a better outcome.更多还原