文章基本信息

标题：Polynomially Bounded Linear Complementarity Problem-based Solvers for Model Predictive Control ⁎
本地全文：下载
作者：Connor R. Stine ; Ambrose A. Adegbege
期刊名称：IFAC PapersOnLine
印刷版ISSN：2405-8963
出版年度：2021
卷号：54
期号：21
页码：67-72
DOI：10.1016/j.ifacol.2021.12.012
语种：English
出版社：Elsevier
摘要：AbstractWe present a computational complexity certificate for a special type of Linear Complementarity problem (LCP) arising in Constrained Linear Quadratic Regulator (LQR) based Model Predictive Control (MPC). By exploiting the special structure of the LCP, we provide sufficient conditions for the LCP to be solvable in a finite number of pivots when processed using Lemke’s complementarity pivoting algorithm. The finite time termination is associated with the existence of a covering vector whose length determines the number of pivot operations needed, and guarantees an upper bound on computational complexity. This result is stronger than earlier assumed in the literature and provides a very promising solution technique for use in embedded MPC applications.
关键词：KeywordsModel Predictive ControlQuadratic ProgrammingLinear Complementarity ProblemConstrained ControlConstrained LQR