主汽温PID参数整定的多目标遗传优化研究

综合考虑舰船锅炉控制系统的跟随性能、扰动抑制性能和鲁棒性能,提出一种基于多目标遗传算法和多属性决策的PID参数设计方法,采用改进非支配解排序的多目标遗传算法求出Pareto最优解。由Pareto最优解构成决策矩阵,使用客观赋权的信息熵法对最优解的属性进行权值计算,再采用逼近理想解的排序方法进行多属性决策研究,对Pareto最优解给出排序。计算某锅炉主汽温PID控制的数值算例,结果表明方法通用性良好,所设计的PID性能优异,适合工程实际应用。

Multi-Objective Genetic Optimization of PID Parameters for Setting Main Steaming Temperature

Based on multi-objective Genetic Algorithm（GA） and multi-attribute decision-making process,a PID parameters design method was developed to obtain an optimum method for setting the main steaming temperature,which involved servo performance,disturbance restraining performance and robustness of ship boiler system.An improved Non-Dominating Sorting Genetic Algorithm II（NSGA II） was employed to calculate the optimum Pareto solutions that determine decision making matrix.Information entropy approach with objective weight was adopted to calculate the attributes weight distribution of the optimum solutions.Then,the study of multi-attribute decision-making（MADM） was carried out by TOPSIS,a sorting method which approximates the ideal solution was obtained for reaching the sorting of optimum Pareto solutions.The calculated results of the main steaming temperature of a boiler show that the method can be widely applied with good performance in practical engineering field.