Linguistic Hedges and Genetic Algorithms on Fuzzy Logic COntroller Design

• Main Authors: Tsao, Ju-Ying, 曹如瑩
• Other Authors: Liu Bin-Da
• Format: Others
• Language: zh-TW
• Published: 1998
• Online Access: http://ndltd.ncl.edu.tw/handle/30660883491682096306

碩士 === 國立成功大學 === 電機工程學系 === 86 === In this thesis, a fuzzy logic controller, called as linguistic hedge fuzzy logiccontroller, designed by the linguistic hedges and genetic algorithms is proposed.The linguistic hedge is a fuzzy operation applied to adjust the membership functionsof the fuzzy sets. If the control results fail to acquire the system requirements, the control objective can still be achieved by adjusting the membership functions of the fuzzy sets or the control rules. The proposed linguistic hedge fuzzy logic controller takes advantage of dynamically modifying the membership functions of a ×3-rule fuzzy logic controller system to achieve a similar performance that a 7×7-rule fuzzy logic controller system possesses. Therefore, the computational complexity is dramatically reduced due to the less number of rules required. Furthermore, a modified genetic algorithm is used to search the optimal linguistic hedge combination. Finally, three well-known control systems ─the truck backer- upper system, the inverted pendulum system, and the nonlinear plant model system ─are used to demonstrate the feasibility of the linguistic hedge fuzzy logic controller.