Application And Reaserch Of Case-based Reasoning Based On Grey-relational Theory In Intelligent Diagnosis System

The fault intelligent diagnosis system, as a type of the artificial intelligence system, has been developing along with artificial intelligence technology. Combining grey-relational analysis and case-based reasoning, a new solution on the fault intelligent diagnosis system for loader is presented. The object-oriented case representation and the hierarchy model of fault cases are investigated deeply. Fault state vector of case is presented, which is additional important information of a fault case. Fault state vector of case set is presented, which is a symbol of the assorted cases. Frequency of fault case is presented, by which the most similar case can be chosen. In addition, the weight vector of a case should be calculated separately according to different case sets, and the fault state vector corresponds with the weight vector. In case retrieval, a retrieval strategy which consists of four steps has been presented, and the four steps are sort choosing, fuzzy choosing, accurate choosing and choosing the best. Furthermore, combining grey-relational theory and Euclidean distance, a new method for computing degree of similarity in CBR is presented, and this method is the core of "accurate choosing". In the two aspects case-based modifying and case-based learning, some new ideas which are in allusion to the problems of traditional methods are presented. In system design and realization, the strategy of combining CBR and RBR in fault intelligent diagnosis system for loader has been investigated deeply. Especially, the realization method of the fault intelligent diagnoses system which is based on CBR is introduced in detail, such as system framework, system functions, diagnosis flow and case database design. Finally, some interfaces and operating steps are given to illustrate case management and knowledge management. At last, the main conclusion is drawn, and research trends in AI are anticipated.