Efficient Solution of Semidefinite Programs for Analysis of Gain Scheduled Controllers
2002 (English)In: Proceedings of the 2002 SIAM Conference on Optimization, 2002Conference paper (Refereed)
Gain scheduling is a very powerful control methodology for systems with time varying parameters. The only requirement is that the process dynamics can be predicted. Often, analysis of a system controlled by again scheduled controller results in solving extremely large optimization problems involving linear matrix inequalities (LMIs) as constraints. Standard algorithms for solving these so called semidefinite programs (SDPs) cannot handle problems of the size commonly encountered in applications. However, the LMIs have a very special structure. If this structure is exploited and is combined with an interior-point method for solving the SDP a very efficient algorithm results.
Place, publisher, year, edition, pages
Engineering and Technology Control Engineering
IdentifiersURN: urn:nbn:se:liu:diva-91138OAI: oai:DiVA.org:liu-91138DiVA: diva2:618396
2002 SIAM Conference on Optimization, Toronto, Canada, May, 2002