Multi-objective optimization based on Genetic Algorithm for PID controller tuning  

作　　者：王国良 阎威武 邵惠鹤 

机构地区：[1]Institute of Automation,Shanghai Jiaotong University

出　　处：《Journal of Harbin Institute of Technology(New Series)》2009年第1期71-74,共4页哈尔滨工业大学学报（英文版）

基　　金：Sponsored by the National Natural Science Foundation of China (Grant No. 60504033)

摘　　要：To get the satisfying performance of a PID controller, this paper presents a novel Pareto-based multi-objective genetic algorithm (MOGA), which can be used to find the appropriate setting of the PID controller by analyzing the pareto optimal surfaces. Rated settings of the controller by two criteria, the error between output and reference signals and control moves, are listed on the pareto surface. Appropriate setting can be chosen under a balance between two criteria for different control purposes. A controller tuning problem for a plant with high order and time delay is chosen as an example. Simulation results show that the method of MOGA is more efficient compared with traditional tuning methods.To get the satisfying performance of a PID controller, this paper presents a novel Pareto - based multi-objective genetic algorithm （MOGA）, which can be used to find the appropriate setting of the PID controller by analyzing the pareto optimal surfaces. Rated settings of the controller by two criteria, the error between output and reference signals and control moves, are listed on the pareto surface. Appropriate setting can be chosen under a balance between two criteria for different control purposes. A controller tuning problem for a plant with high order and time delay is chosen as an example. Simulation results show that the method of MOGA is more efficient compared with traditional tuning methods.

关 键 词：multi-objective optimization genetic algorithms PID controller 

分 类 号：O224[理学—运筹学与控制论] TP273[理学—数学]