基于遗传算法的模糊神经网络在动态系统辨识中的应用

摘要

复杂不规则系统的语言建模构成了许多控制／决策系统的核心问题，模糊逻辑是进行语言建模最有效的方法之一。本文介绍了一种基于模糊逻辑、神经网络和遗传算法的语言建模方法，并给出了新型的混合学习算法，即：首先由自组织算法确定出模糊神经网络的初始隶属度函数；其次由最大匹配因子学习算法完成模糊规则确定；最后提出了一种改进的遗传算法用来优化调节已经获得的隶属度函数。通过具体的仿真实例说明了所提出的建模方法在动态系统辨识中的有效性。%A method for linguistic modeling based on fuzzy logic, neuralnetworks and genetic algorithm is introduced in this paper, and a new hybrid learning algorithm is proposed: Firstly,the initial membership functions of the fuzzy neural network are found by using the self-organization feature map algorithm; then the maximum matching-factor algorithm is used to determine the fuzzy rules; finally, a modified genetic algorithm is presented and used for optimized tuning of the membership functions obtained. A simulation example demonstrates the efficiency of the proposed scheme.