摘要
模糊PI控制器具有鲁棒性强、控制灵活等优点,但是将其应用于纯迟延系统时超调量较大、响应速度慢。针对此提出了一种基于遗传算法的模糊PI控制器,使用遗传算法对模糊逻辑系统参数进行训练。在以往的模糊逻辑系统建立过程中,主要依靠专家知识或工作人员经验来确定其主要参数(如模糊推理规则和隶属函数参数等),而该文利用遗传算法对样本数据进行优化来获取系统参数。在遗传算法中,将推理规则和隶属函数参数的确定结合在一起,从而确定最优的模糊逻辑系统。仿真试验结果表明,由该方法得到的控制器用于纯迟延系统具有响应快,超调量小等优点。
Fuzzy PI controller has good robustness and flexible control, but it has maximal overshoot and slow response when used in the process with large time delay. To overcome the defect a new kind of fuzzy PI controller based on genetic algorithm is introduced, which uses genetic algorithm to train the parameters of fuzzy logic system. Traditional methods for choosing fuzzy rules and membership functions mainly depend on the knowledge and experience from experts and technologic faculty. In this paper, however, genetic algorithm is used to obtain the parameters of system. In the procedure of genetic algorithm, fuzzy rules and membership functions are combined together to obtain optimal fuzzy logic system. Result of simulation test for process with large time delay shows minimal overshoot and a good rising time and settling time.
出处
《计算机仿真》
CSCD
2006年第5期153-155,共3页
Computer Simulation
关键词
模糊比例积分控制器
遗传算法
模糊逻辑系统
Fuzzy proportional- integral controller
Genetic algorithm
Fuzzy logic system
