CMAC最优设计及其算法——GA技术优化CMAC偏移矢量分布

采用GA(Genetic Algorithm)技术实现CMAC(cerebellar Model Articulation Controller)最优设计及算法.该方法解决了CMAC与其学习对象的整体优化问题,具有理论 意义和实用价值.仿真结果证明该方法是成功的和有效的.对不同的客观对象(如空间曲面), 可以采用GA技术找到CMAc的最优内部表示(偏移矢量分布),实现一般CMAC难以达到 的精度.该方法比Albus的CMAC和Parks等的CMAC学习效果都有不同程度的提高,适 合于要求高精度学习的情况.同时给出了任意偏移矢量分布的CMAC算法.

Optimal Design and Algorithm of CMAC-Optimization of CMAC Displacement Vectors Distribution by Genetic Algorithm

Optimal design and algorithm of CMAC are implemented by genetic algorithm (GA) in this paper. Such a method has solved the optimization problem considering both CMAC and its learing objects. Simulations show that the proposed method is successful and effective. For different objects learned (such as space planes), different internal presentations (displacement vectors distribution)of CMAC can be found by GA, so that the optimal CMAC possesses higher learning precision than the general CMAC can reach. The proposed method has some improvements on both Albus' CMAC and Parks' CMAC. This method is suitable to the high precision learning problem. The CMAC algorithm for any displacement vectors distribution is also proposed.