优先级多目标稳定化约束模型预测控制

Stabilizing constrained model predictive control with prioritized multiobjectives

针对目标函数的不同优先级问题,提出一种约束多变量线性定常系统的稳定化多目标模型预测控制策略.首先,基于多目标优化理论给出多目标预测控制问题的字典序最优解结果,并在此基础上考虑目标函数的优先级,重新将多目标预测控制问题定义为字典序多目标预测控制问题;然后,采用终端约束、终端罚函数和局部状态反馈律等三要素,证明多目标预测控制闭环系统是渐近稳定的;最后,通过一个仿真实例验证了所提出方法的有效性.

For the different pdoritization problem of objective functions, a multi-objectives model predictive control scheme with guaranteed stability is proposed for constrained multi-variables linear invariant-time systems. Firstly, some results on lexicographic optimal solutions on the multi-objectives predictive control problem are given based on the multi-objective optimization theory. Then the multi-objectives predictive control problem is reformulated as the lexicographic one that is able to deal with the different priodtization of objective functions. Then, ingredients of terminal constraints, terminal penalty functions and local state feedback laws are used to prove the asymptotical stability of the multi-objectives predictive control closed-loop system. Finally, a simulation example is exploited to illustrate the effectiveness of the results obtained.