论文部分内容阅读
提出一种基于遗传算法的大惯性系统辨识方法。该方法将动态补偿和遗传算法相结合 ,实现了系统的参数辨识。为避免遗传早熟而进入局部最优 ,对交换和变异操作进行了改进 ,提出一种防止近亲繁殖的交换策略 ,在不明显增加基因操作计算量的前提下 ,有效地避免了基因缺失。仿真结果表明 ,该方法对过阻尼系统的辨识是有效的。
A large inertial system identification method based on genetic algorithm is proposed. The method combines dynamic compensation and genetic algorithm to realize the system parameter identification. In order to avoid genetic precocity and enter the local optimum, the exchange and mutation operations have been improved, and a strategy for preventing inbreeding is proposed. The gene deletion can be effectively avoided without significantly increasing the computational cost of genetic manipulation. Simulation results show that this method is effective in identifying over-damped systems.