摘要
针对开放环境中多Agent合作诊断问题,本文提出了一种新的、基于TMS的冲突消解策略。在文中描述的信念知识表示基础上,研究了利用异构的Agent对同一对象进行诊断,以及诊断结果间冲突的类型和信念修正方法。该系统不同于简单的投票机制,既可以纠正错误的假设,提高诊断结果的可靠性,又可以实现知识库的一致性维护。同时也作了形式化的分析和描述。
In order to solve the conflict resolution problem existed in multi-agent cooperative diagnosis in open environments, a new strategy based on TMS is presented. The types of result conflict among agents and the methods of belief revision are studied based on the belief knowledge representation described in this paper. This mechanism can not only enhance reliability of diagnosis and correct the wrong hypotheses, but also realize consistency of agents diagnosis by maintaining knowledge base. The formal analysis and description are presented to support the method.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2000年第4期395-399,共5页
Pattern Recognition and Artificial Intelligence
基金
国家863-306课题资助项目
