采用免疫进化算法优化设计径向基函数模糊神经网络控制器

基于生物免疫系统的计算智能近年来正逐渐成为一个研究热点.针对模糊神经网络控制器难于设计的问题,提出了一种免疫进化算法用于径向基函数模糊神经网络控制器参数的优化设计.首先将控制器参数进行编码表示成个体,并由若干随机个体组成初始群体;然后模拟生物适应性免疫应答过程,通过扩展操作在群体中较优秀个体的小邻域内进行局部搜索,同时利用突变操作在较差个体的大邻域内搜索;最后将设计的控制器用于控制倒立摆系统,仿真结果验证了该控制器的有效性.

Optimal design of radial basis function fuzzy neural network controllerbased on immune evolutionary algorithm

The computational intelligence based on biology immune system is becoming the focal point in recent years.Aiming at the design difficulty for fuzzy neural network controller,an immune evolutionary algorithm is proposed to design the parameters of a radial basis function fuzzy neural network controller.First,the parameters of the controller were encoded into an individual,and the initial population was composed of some random individuals; then,simulating the process of biology adaptive immune response,expansion operation was used to perform local searching in a small neighborhood of the better individuals in population,and mutation operation was used to search in a large neighborhood of the worse individuals.Finally,the designed controller was employed to control an inverted pendulum system,and the simulation results verified the effectiveness of the controller.