Polynomial Interior Point Algorithms for General Linear Complementarity Problems

Tibor Illés, Marianna Nagy, Tamás Terlaky

Abstract


Linear Complementarity Problems (LCPs) belong to the class of NP-complete problems. Therefore we can not expect a polynomial time solution method for LCPs without requiring some special property of the coefficient matrix. Following our recently published ideas we generalize affine scaling and predictor-corrector interior point algorithms to solve LCPs with general matrices in EP-sense, namely, our generalized interior point algorithms either solve the problems with rational coefficient matrix in polynomial time or give a polynomial size certificate that our matrix does not belong to the set of P * (~κ) matrices, with arbitrary large, but apriori fixed, rational, positive ~κ.

Keywords


linear complementarity problem, sufficient matrix, $\mathcal{P}_*$-matrix, interior point method, affine scaling method, predictor-corrector algorithm

Full Text:

PDF


Algorithmic Operations Research. ISSN: 1718-3235