Case-based Fuzzy Expert System

Case-Based Reasoning (CBR) is an advance reasoning technology in AI and ES. Since Roser Schank wrote 'Dynamic Memory' in 1982, CBR has made important progress in theory and application. CBR system poses many features, such as, memory accumulation , quick index , march ,' learn and induction. CBR can solve many difficult problems that traditional handling knowledge system can't solve, for instance, knowledge handling efficiency low outside system 'boundary', march conflict, combination explosion, self-learn difficult, complex problem class difficult and train taking long time in artificial nerve net.Many respects in CBR expert system still be researched, such as, case representation, case retrain , case index, case reuse, case revise and case retain. Especially, it is not reported how the above-mentioned respects induce and research the fuzzy information in case.The article studies the case-based fuzzy expert system from theory and application overall. Chapter 1 states the meaning, resource and background of the project, and summaries the reference conferences. Chapter 2 presents a total frame construction to research and develop the CBR system. Chapter 3 studies the fuzzy case representation, and presents a knowledge representation method of oriented object. Combining the knowledge representation, the article put forward to a data construction organization form of case base. Chapter 4 studies the fuzzy case retrain. Chapter 5 studies the fuzzy case renew and learn. Chapter 6 studies the fuzzy case reasoning and presents three methods of determining case feature weights, that is the weight-determined method combining Delphi and AHP, the quadratic programming weight-determined method and the credit statistics weight-determined method. Meanwhile, the article presents a case synthesis reasoning method in theory. According to the achievements obtained in theory research above chapters and the practice problem in the quality control management system at Silu copper corporation, and applying artificial nerve net to case-based reasoning, Chapter 7 developed the electrolytic copper quality control counsel expert system. The developed soft poses the features of operation convenient and function perfect.