基于EV-SPEA2的过热汽温控制系统优化方法  

OPTIMISATION OF SUPERHEATED STEAM TEMPERATURE CONTROL SYSTEM BASED ON EV-SPEA2

作　　者：周霞[1,2] 

机构地区：[1]金陵科技学院机电工程学院,江苏南京211169 [2]东南大学能源与环境学院,江苏南京210096

出　　处：《计算机应用与软件》2015年第10期260-264,301,共6页Computer Applications and Software

基　　金：国家自然科学基金项目(51306082);金陵科技学院博士启动基金项目(JIT-B-201218)

摘　　要：为将决策者的期望信息融入多目标优化问题的求解过程,提出一种基于期望值的偏好距离计算方法。基于此改进了强度多目标优化算法的适应度计算方法与环境选择策略,并提出了算法EV-SPEA2。针对经典多目标优化函数算例的验证表明,新算法可以根据决策者的期望信息,提供大量距离期望值较近的解。将EV-SPEA2应用于锅炉过热汽温控制系统PID参数整定问题,仿真结果表明,运行EV-SPEA2可以获得多组位于Pareto前端,且符合决策者偏好的控制器参数。In order to integrate the expected information of decision maker into the procedure of solution in regard to multi-objective optimisation,we proposed an expected value-based preference distance calculation method,and improved according to it the fitness assignment method and environment selection strategy of strength multi-objective optimisation algorithm,as well as proposed the algorithm of EV-SPEA2.It was demonstrated by the verification for the example of classic multi-objective optimisation function that the EV-SPEA2 was able to provide a great deal of solutions approaching to the expected value according to the expected information of decision makers.Applying the EV-SPEA2 in PID controller parameter tuning in boiler superheated steam temperature cascade control system,simulation results showed that quite a few sets of controller parameters at the front of Pareto and in accordance with the preference of decision maker could be achieved by running the EV-SPEA2.

关 键 词：多目标优化 适应度计算 环境选择 期望值 过热汽温控制系统 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]