A Hybrid GA/CMAC Algorithm for Improving the Accuracy in Modeling of Function Approximation
The structure of the Cerebellar Model Articulation Controller (CMAC) is based on the operation of the human cerebellum with good characteristics of generalization and fast learning property. Owing to these attractive properties,the CMAC can be used to various application areas, such as approximate nonlinear functions, pattern recognition, and so on. In this paper, Genetic Algorithms is employed on the training of CMAC, which can optimize parameters of the CMAC. To demonstrate the effectiveness of the proposed method, simulation experiment results were illustrated. The results show that, in function approximation, the proposed strategy has better accuracy than the conventional CMAC.
CMAC Neural Networks Genetic Algorithms Approximate Optimize
Xin Tan Huaqian Yang
School of Communication, Chongqing University of Posts and Telecommunications Chongqing 400065, Chin Department of Computer and Modern Education Technology, Chongqing Education College Chongqing 400067
国际会议
杭州
英文
430-432
2006-10-12（万方平台首次上网日期，不代表论文的发表时间）