论文部分内容阅读
该文首先建造了一种包括多诊断Agent的诊断系统结构。为提高诊断A gent求解效率 ,对诊断Agent内部的故障模式进行聚类 ,将故障模式聚集成多个故障原型 ,使诊断只在特定空间上进行求解 ,避免搜索过程的盲目性和无关性 ;在此基础上 ,设计了一种基于故障原型的诊断Agent内部序贯诊断算法。另外考虑到诊断对象固有的关联性特征造成的多Agent间的诊断耦合 ,提出了以故障原型为依据建立Agent间关联模型的方法。基于故障原型的序贯诊断算法和Agent关联模型使Agent的内部诊断和Agent间的协作行为得以有效地联系。文中最后通过一个化工过程诊断实例对提出的多Agent诊断求解方法进行了验证。
This paper first builds a diagnostic system structure including multi-diagnosis Agent. In order to improve the efficiency of diagnosing A gent, clustering the fault patterns inside the diagnostic agent and clustering the fault patterns into multiple fault prototypes, the diagnosis is only solved in a specific space to avoid the blindness and irrelevance of the search process; in Based on this, a kind of internal diagnostic algorithm based on fault prototype is designed. In addition, the diagnostic coupling between multi-agents, which is caused by the inherent relevance of diagnostic objects, is proposed. A method of establishing association model between agents based on fault prototype is proposed. The sequential diagnosis algorithm based on fault prototype and the agent association model make the internal diagnosis of Agent and the collaboration between Agents effectively connected. In the end, the method of solving multi-agent diagnosis is verified through a case of chemical process diagnosis.