不确定系统的鲁棒与随机模型预测控制算法比较研究

A Comparative Study on Algorithms of Robust and Stochastic MPC for Uncertain Systems

近几十年来,不确定系统模型预测控制的理论和应用得到了飞速发展.本文简要地回顾了不确定系统中鲁棒模型预测控制和随机模型预测控制的发展历史,总结了它们的相关应用,并较为细致地分析了线性不确定系统模型预测控制的各种主要算法.通过总结各种算法的通用模型、运作方式、问题规模,以及它们保证递归可行性、稳定性的方法,分析了部分算法可行域间的关系,揭示了各种算法的主要特点、适用场合和未来可发展方向,并通过仿真实例直观地分析了各种算法的性能和可靠性.

In recent years, the development of model predictive control（MPC） of uncertain systems has been remarkable.This paper briefly reviews the development of robust MPC and stochastic MPC, summarizes their applications, and expounds and discusses the main algorithms of linear uncertain systems in these two fields. By summarizing their general models, the ways they work, computational complexities, and the ideas they use to ensure recursive feasibility and stability,we reveal the relationship of feasible sets among some of them and unravel the main features of these algorithms and their application situations. Finally, we demonstrate the performance of all the algorithms through certain simulation cases and give some indication on the future development of these two fields.