A Penalty Technique for Nonlinear Complementarity Problems

Abstract

In this paper, we first give a new equivalent optimization form to nonlinear complementarity problems and then establish a damped Newton method in which penalty technique is used. The subproblems of the method are lower-dimensional linear complementarity problems. We prove that the algorithm converges globally for strongly monotone complementarity problems. Under certain conditions, the method possesses quadratic convergence. Few numerical results are also reported.