New predictor-corrector interior-point algorithm for sufficient linear complementarity problemsOptimization Seminar of the Corvinus Center for Operational Research
Talk at the Optimization Seminar of the Corvinus Center for Operational Research:
Dr. Petra Renáta Rigó (CIAS – CCOR, CUB) New predictor-corrector interior-point algorithm for sufficient linear complementarity problems.Joint work with Zsolt Darvay and Tibor Illés
Abstract: We present a newpredictor-corrector (PC) interior-point algorithm (IPA) for solving sufficientlinear complementarity problems (LCPs). The introduced IPA uses a new type ofalgebraic equivalent transformation (AET) on the centering equations of thesystem defining the central path. We apply the square root function in this newtype of AET in order to determine the search directions. We prove that the PCIPA retains polynomial iteration complexity in the handicap of the problem’smatrix, the size of the problem and the bit size of the data.
Organizer: CorvinusInstitute for Advanced Studies – Corvinus Centre for Operations Research, Dr.Tibor Illés