基于遗传算法的模糊逻辑控制器优化设计

设计了一种新的用于模糊逻辑控制器优化的遗传算法，实现了模糊逻辑控制器参数和结构的同步快速优化。该算法对经典的简单遗传算法进行了改良，设计了一种带区间限制的十进制编码方案，实现了模糊控制器隶属度函数和控制规则的联合编码；并设计了受限分步一致交叉和变异算子，改进了经典赌轮选择法,避免了病态个体产生，加快了收敛速度。通过对典型的一阶和二阶工业对象控制器优化的仿真研究表明，相对于一般的遗传算法，该算法在群体规模较小（20或40）的情况下，能够在20代左右快速收敛到理想结果，并且在进化过程中很少产生病态个体，因此能够在小群体规模下，实现模糊控制器参数和结构的快速同步优化。

Optimization of Fuzzy Logic Controller Based on Genetic Algorithm

A novel genetic algorithm(GA) which fits to optimization of fuzzy logic controllers(FLC) is proposed. The GA can optimize parameters and structure of FLC simultaneously. The algorithm refined traditional simple genetic algorithm. A new decimal coding schema was designed to implement the unified coding of member functions and rules of FLC. An improved selection method, a constrained and stepped crossover and mutation operator are presented to decrease the illed items and speed up the convergence.The algorithm is applied to the optimization of FLCs for typical first order and second order plant.Simulations show that the algorithm can converge to good result through about 20 generation evolutions and generate fewer illed items at smaller population. Therefore,the new GA can quickly optimize the parameters and structure of FLC for smaller initial population.