A New Interior Point Method for Linear Complementarity Problem
Résumé
For a given n-vector q and a real square matrix M∈IR^{n×n}, the linear complementarity problem, denoted LCP(M,q), is that of finding nonnegative vector z∈IRⁿ such that z^{T}(Mz+q)=0 and Mz+q≥0. In this paper we suppose that the matrix M must be a symmetric and positive definite and the set S={z∈ IRⁿ / z>0 and Mz+q>0}; named interior points set of the LCP(M,q) must be nonempty. The aim of this paper is to show that the LCP(M,q) is completely equivalent to a convex quadratic programming problem (CQPP) under linear constraints. To solve the second problem, we propose an iterative method of interior points which converge in polynomial time to the exact solution; this convergence requires at most o(n^{0,5}L) iterations, where n is the number of the variables and L is the length of a binary coding of the input; furthermore, the algorithm does not exceed o(n^{3,5}L) arithmetic operations until its convergence and in the end, we close our paper with some numerical examples which illustrate our theoretical results.
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...