可重构制造系统故障诊断多Agent自学习模型  

Fault diagnosis with self-learning for reconfigurable manufacturing system based on multi-agent system

作　　者：贺鹏[1] 王庆林[2] 

机构地区：[1]上海工程技术大学城市轨道交通学院,上海201620 [2]北京理工大学信息科学技术学院,北京100081

出　　处：《计算机工程与设计》2007年第8期1741-1743,1791,共4页Computer Engineering and Design

基　　金：国家973重点基础研究发展计划基金项目(2002CB312204)

摘　　要：提出了一种基于多Agent的可重构制造系统故障诊断自学习模型,模型为3层结构,从下向上分别为监测Agent层、诊断Agent层和管理Agent层,详细分析了模型的故障诊断自学习机理,自学习性是基于数据挖掘和基于粗糙集的。示例表明可重构制造系统重构前后,均能有效完成故障诊断。The reconfigurable manufacturing system （RMS） need better fault diagnosis. Firstly, a fault diagnose model with selflearning for RMS based on multi-agent system （MAS） is proposed. The model consists of three levels including monitor agent level, diagnosis agent level and manager agent level. The fault diagnosis mechanism of the model is analyzed. Then, the self-learning mechanism in the fault diagnosis model is discussed in detail. The self-learning mechanism includes the method of data mining and the method of rough set. Finally, an example is used to test the model and the result expresses that the model is feasible before and after reconfiguration.

关 键 词：可重构制造系统 故障诊断 多AGENT系统 数据挖掘 粗糙集 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]