基于免疫算法自调节变异率的FNNC参数优化及其应用  

The Parameters Optimization of FNNC Based on Adaptive Mutation of Immune Algorithm and Its Application

作　　者：廖迎新 吴敏[1] 

机构地区：[1]中南大学信息科学与工程学院

出　　处：《系统工程》2006年第3期110-113,共4页Systems Engineering

基　　金：国家杰出青年科学基金资助项目(60425310);湖南省科技计划资助项目(04FJ3029)

摘　　要：针对复杂钢坯加热过程,提出了一种自调节变异率的免疫进化模糊神经网络控制(IE-FNNC)算法。首先根据现场样本数据建立过程神经网络模型;然后基于该模型,采用规则优化算法,确定模糊神经网络控制器(FNNC)的最佳规则数;最后由FNNC的规则优化所得参数构造初始种群的一个解,采用自调节变异率的免疫进化(IE)算法对FNNC参数优化。该算法具有全局寻优和局部求精能力,仿真结果证实了其有效性。The immune evolutionary （IE） algorithm based on self-adaptive mutation is used to optimize the parameters of the fuzzy neural network eontroller（FNNC） for the complex billet heating process. First, the neural model of the process is set up via the observation data;then, based on the model, the rule optimization algorithm is introduced to obtain the optimal rule numbers of the FNNC. Finally, combining the FNNC parameters obtained with the rule optimization into an individual in the initial population, the IE algorithm based on self-adaptive mutation is adopted to more optimize the parameters of the FNNC. The proposed search strategy is a global and precise local optimization algorithm. Simulations show the effectiveness of the constructed system.

关 键 词：FNNC 规则优化 免疫进化 加热炉 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]