A Decomposition Approach for Solving KYP-SDPs

Authors:	Wallin Ragnar, Linköpings universitet, Sweden Kao Chung-Yao, University of Melbourne, Australia Hansson Anders, Linköpings universitet, Sweden
Topic:	2.5 Robust Control
Session:	Positive Polynomials and LMI Optimization with Applications to Robust Control
Keywords:	Optimization, Decomposition Methods, Robust Control

Abstract

Semidefinite programs originating from the Kalman-Yakubovich-Popovlemma are convex optimization problems and there exist polynomialtime algorithms that solve them. However, the number of variables isoften very large making the computational time extremely long.Algorithms more efficient than general purpose solvers are thusneeded. In this paper a generalized Benders decomposition algorithmis applied to the problem to improve efficiency.