| Summary: | 碩士 === 義守大學 === 工業工程與管理學系 === 92 === During the last decades, the six generation systems (SGS) that included fuzzy logic, neural networks and genetic algorithms have been highly developed. In addition, there are many investigates on the merging of fuzzy logic, neural networks and genetic algorithms in control, expert systems or other fields. In this thesis, a new heuristic algorithm (electromagnetism-like mechanism, EM) for global optimization is introduced. Electromagnetism-like mechanism simulates the electromagnetism theory of physics by considering each sample point to be an electrical charge. The algorithm utilizes the attraction-repulsion mechanism to move the sample points towards the optimum. Besides, the electromagnetism-like mechanism is not easily falling into local optimum. Therefore, the purpose of this study is using the electromagnetism-like mechanism to develop an electromagnetism-like mechanism based fuzzy neural networks (EMFNN), and employ this EMFNN to train fuzzy IF-THEN rules. The EMFNN also compares with conventional fuzzy neural networks (FNN). According to the examples and cases, the EMFNN could decrease the network error and we could find the EMFNN have better learning capability than conventional fuzzy neural networks, and it also could successfully generalize new fuzzy if-then rules. Thus it can be known that the training results of EMFNN in our cases are better than that of conventional fuzzy neural networks.
|
|---|
