复杂系统的MAS分布式协作求解研究

在传统故障诊断基础上,将多Agent技术应用于动态、分布、实时和不确定的复杂系统故障诊断领域，是求解复杂过程故障诊断问题的一种新尝试。讨论了基于MAS的分布武智能故障诊断方法和过程。基于一种分布式Agent诊断系统结构原型系统，重点研究了诊断问题的多Agent间的交互、协作以及诊断决策问题，提出了分布式信息交换和分布武诊断的BDI结构的推理机模型，并用CBR技术对其实现了协作求解。应用表明，该系统能快速、准确地进行故障成因分析．取得了与专家相似的诊断结果。

Research on Complex System Intelligence Fault Diagnosis Based on Multi-agent

Modern fault diagnosis is a dynamic, distributed, flexible, real-time and uncertain complex system. The paper tried to introduce multi-agent technology (MAS) to the complicated fault diagnosis. The paper discusses the methods and process of the distributed intelligence fault diagnosis based on MAS. A systematical system or prototype system based MAS are put forward. The task decomposition, distribution, inner diagnosis mechanism, multi-agent mutation, cooperation, and correlation model of the fault diagnosis problem is also discussed. Diagnosis system can be divided into applied agent and service agent, using distribute communication mechanism to realize intercommunication of information. it realize distribute system's information exchange and diagnosis, and enhance system's efficiency, security for problem-dealing. When applied on monitor and controlling system of electric enterprise, the model can offer us a quick and accurate analysis, and also give us reasonable, constructive and decisive advice.