改进的自适应遗传算法在控制器PID参数优化中的应用

针对遗传算法(GA)中自适应机制存在理论性和可行性之间的矛盾,提出一种行之有效的自适应算法。该算法根据个体和群体的适应度来确定个体的交叉变异概率,并据此采用无放回抽样的方式进行交叉变异。用一个非常复杂的数学函数对新算法进行了测试,结果表明改进算法克服了传统GA难以解决的早熟和局部收敛的问题,并且收敛速度大有提高。在此基础上将其应用于锅炉过热汽温串级控制系统的PID参数优化,仿真结果表明了其优越性,具有很好的应用价值。

An Improved Adaptive Genetic Algorithm and Its Application in Parameter Optimization of PID Controller

Aiming at the contradiction between theory and feasibility existed in adaptive mechanism of the traditional genetic algorithm(GA),this paper presents an improved algorithm.The new algorithm computes crossover and mutation probability according to the fitness,and hereby carries on crossover and mutation by the way of sampling without replacing.It is tested with a complex mathematics function,the experimental results show that it remedies the prematurity and local convergence,and convergence velocity increases prominently.The improved algorithm is applied to PID controller parameter optimization on boiler's superheated steam temperature.The simulation result shows it is valid,and has application values.