抄録
We have already reported about the case-based fault restoration support system in erectrical power systems. Case-based reasoning systems can improve the performance of problem solving by storing all new cases into the case-base. However, the performance of retrieval declines because of the expansion of the case-base. Generalizing cases and merging similar cases prevent from declining the retrieval performance. Besides, it is easy for users to understand the applicable scope of the cases. Cases also need to be specialized to adapt their applicable scope to the environment. This means that cases which were useful for solving problems in the past often become unsuitable since environmental conditions would change as time passes. For example, after new substations or electric power facilities have been built, some cases may become unable to be applied to current failures. This paper describes refinement methods of case-base by generalizing and specializing cases. The experimental results through computer simulations demonstrated that both the number of cases and the time for problem solving decreased to around 5%, and the percentage of getting optimum solutions using a case with the highest priority increased, compared to a case where all new cases are added to the case base.
