基于三阶染色体结构的自适应进化规划及应用

Self-adaptive Evolutionary-programming Based on Three Layer Chromosome and Application

针对模糊辨识中规则数多少难以选择、输入量在各规则中重要程度难以确定以及模糊规则参数难以优化等问题,模拟生物基因之间相互联系的特点,用三阶染色体结构对完全模糊规则基进行描述。根据每代不同个体的适应度信息获得不同个体的变异率,采用自适应进化规划算法,实现模糊规则基各级染色体引导优化,克服了用试凑方法确定参数对辨识精度的消极影响,算法的复杂程度大大减小,为模糊建模提供了一种新颖的方案。通过对两类不同非线性系统的仿真,证明了此方法的有效性。

A new method which imitates features of relating in biologys gene is proposed in the paper, and it can solve the problems of fuzzy identify in nonlinear system, such as difficulty in choosing number of rules, weight of input variable in different rules and difficult in optimizing parameters of fuzzy rule bases, etc. In the paper, a three layers chromosome is used to describe perfect rule bases. Different to old methods, an adaptive evolutionary algorithm is used to realize guiding optimum based on the fitness information of different individual. Moreover, the whole chromosome is optimized simultaneously. The new method remedies the bad effect of compromise proposal in identifying of old method and the given algorithm is simple. The effectiveness of the given method is demonstrated by simulations of different nonlinear systems.