PID参数优化的实值遗传算法  

Optimization of PID Parameters Based on Real-Valued Genetic Algorithm

作　　者：陈华艳[1] 周文彬[1] 

机构地区：[1]华南理工大学自动化科学与工程学院,广东广州510640

出　　处：《兵工自动化》2004年第5期59-60,共2页Ordnance Industry Automation

摘　　要：基于实值遗传算法的PID参数优化采用实数编码,通过参数实值编码初始化种群,计算适应度函数适配值。经对寻优的比例系数、积分和微分时间常数三个参数实数编码,根据适配值的大小进行遗传选择,将复制产生的个体进行交叉和变异,产生新种群。以迭代100次为算法终止条件,输出最优PID参数。对二阶延迟系统Matlab仿真,证明该算法优越。PID parameter based on real-valued genetic algorithm was optimized with real number coding, Original colony was initialized through parameter real-valued coding, adaptive value of function was calculated. Parameters of proportion factor, integral and differential time constants were coded. Individual was selected according to big and small of adaptive value, and individual copied was exchanged to create the next generation colony. Taking iterative 100 times as end the evolution, optimum PID parameter was outputted. The two-delay system was simulated with Matlab, simulation result shows that the method is superior to the simple genetic algorithm.

关 键 词：实值遗传算法 PID参数 优化 

分 类 号：TP273.1[自动化与计算机技术—检测技术与自动化装置]