双层结构模型预测控制稳态优化层QP与LP求解模态分析  

Modal Analysis of QP and LP Solution in Steady-state Optimization Layer of Two-layer Structure MPC

作　　者：王洪瑞 张鑫[1,2,3] 刘博文[1,2,3,4] 蔺娜[5] 邹涛 WANG Hongrui;ZHANG Xin;LIU Bowen;LIN Na;ZOU Tao(Key Laboratory of Networked Control Systems,Chinese Academy of Sciences,Shenyang,Liaoning 110016,China;Shenyang Institute of Automation,Chinese Academy of Sciences,Shenyang,Liaoning 110016,China;Institutes for Robotics and Intelligent Manufacturing,Chinese Academy of Sciences,Shenyang,Liaoning 110169,China;University of Chinese Academy of Sciences,Beijing 100049,China;National Science and Technology Venture Capital Development Center,Beijing 100038,China;Guangzhou University,Guangzhou,Guangdong 510006,China)

机构地区：[1]中国科学院网络化控制系统重点实验室,辽宁沈阳110016 [2]中国科学院沈阳自动化研究所,辽宁沈阳110016 [3]中国科学院机器人与智能制造创新研究院,辽宁沈阳110169 [4]中国科学院大学,北京100049 [5]国家科技风险开发事业中心,北京100038 [6]广州大学,广东广州510006

出　　处：《控制与信息技术》2020年第2期1-7,11,共8页CONTROL AND INFORMATION TECHNOLOGY

基　　金：国家重点研发计划(2018YFC0604403);国家自然科学基金项目(61773366);辽宁省博士启动基金(20180540066);辽宁省自然科学基金资助计划项目(2019-KF-03-07)。

摘　　要：为解决双层结构预测控制在工程应用中存在的通用化需求及参数配置问题,针对稳态优化层研究两阶段求解策略的标准化过程,形成一系列线性规划(LP)和二次规划(QP)问题;提出目标函数含绝对值及数值截断误差问题的解决策略;针对所形成的LP及QP问题,从几何关系及代数关系两方面分析了解其特性,并分析了模型失配对LP及QP解的影响。仿真验证结果表明,随着决策变量权重系数变化,QP比LP的解具有更好的连续性,QP的权重系数能更精确调节决策变量的偏向性;可行域边界的变化导致LP的解一定发生变化,而一直在约束空间内部的QP解则不随时间变化。In order to solve the problems of generalized demand and parameter configuration in the engineering application of two-layer predictive control,a series of linear programming(LP)or quadratic programming(QP)problems are formed by studying the standardization process of the two-stage solution strategy in the steady-state optimization layer.It proposed a solution strategy of the objective function including absolute value and numerical truncation error,and analyzed the characteristics of solutions from the geometric and algebraic relations aiming at the problems of LP and QP.Simulation results prove that QP had better continuity than LP when the weight coefficient of decision variables are changed,and the weight coefficient of QP can more accurately adjust the bias of decision variables.The change of feasible region boundary results in the solution change of LP while the solution of QP which has been in the constraint space does not change with time.

关 键 词：模型预测控制 稳态优化 双层结构 二次规划(QP) 线性规划(LP) 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]