| Summary: | 碩士 === 國立中正大學 === 電機工程所 === 93 === A case-based reasoning (CBR) system uses the past experience of problem solving to solve a new problem. CBR has successful examples in many domains, such as diagnosis, analysis, planning, design, and decision supporting. The past problem is similar to the new problem but not the same. The past experience of solving problem is not completely fit for the new problem. Thus, past experience that solved the problem must be adapted to solve the new problem. Case adaptation is a challenging task in CBR. We propose a method of adaptation operators, and this method is applied into the process of the case adaptation in the hierarchical case-based reasoning (HCBR). The objective of the method is to reduce the complexity of case adaptation, and to accomplish modification of the solution part. All cases are organized in the manner of the hierarchy, and cases can be decomposed into abstract cases and concrete cases. We can substitute the solution part of the abstract case in the higher levels of the hierarchy, and this introduces the deletion and addition of concrete cases in the lower levels of the hierarchy. In this thesis, we apply the adaptation operators to the retrieval of execution steps. Through using our approach, we can adapt the execution step of the old goal, and the adapted execution step can achieve the new goal.
|
|---|
