A two-step smoothing Levenberg-Marquardt method for
nonlinear complementarity problems
Linsen Songa,*, Yan Gaob
ABSTRACT: By reformulating the nonlinear complementarity problem into a system of nonsmooth equations, we
proposed a two-step smoothing Levenberg-Marquardt method with global convergence in this paper, where not only
one step but also an approximate step is computed at each iteration. Under local error bound condition, the local
convergence result of the present algorithm is also obtained. Finally, numerical experiments illustrate the effectiveness
of the present method when compare it with one-step smoothing Levenberg-Marquardt methods.