2020/12/08 by Goran Banjac, Banjac, Goran, Jianzhe Zhen +5
Engineering · Mathematics · #Advanced Control Systems Optimization #Advanced Optimization Algorithms Research #FOS: Mathematics #Optimization and Control (math.OC) #Stability and Control of Uncertain Systems
paper · pdf · doi:10.48550/arxiv.2012.04688
openalex publication_date 2020/12/08 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Various control schemes rely on a solution of a convex optimization problem\ninvolving a particular robust quadratic constraint, which can be reformulated\nas a linear matrix inequality using the well-known \S-lemma.\nHowever, the computational effort required to solve the resulting semidefinite\nprogram may be prohibitively large for real-time applications requiring a\nrepeated solution of such a problem. We use some recent advances in robust\noptimization that allow us to reformulate such a robust constraint as a set of\nlinear and second-order cone constraints, which are computationally better\nsuited to real-time applications. A numerical example demonstrates a huge\nspeedup that can be obtained using the proposed reformulation.\n