One-step smoothing Newton method for solving the mixed complementarity problem with a \(P_{0}\) function (Q1039706)

In the numerical solution of the mixed complementarity problem, a smoothing method uses a smoothing function to approximate the mixed complementarity problem via a family of parametrized smooth equations and solves the smooth equations approximately at each iteration. This paper presents a new smoothing function which is obtained by smoothing the perturbed mid function. The new smoothing function has a property which allows the study of the boundedness of the iteration sequence generated by the smoothing methods for solving the mixed complementarity problem with a \(P_0\) function under the assumption the solution set of the mixed complementarity problem is nonempty and bounded. A one-step smoothing Newton algorithm is presented. The global convergence and local superlinear convergence of the algorithm under certain conditions are established. Some numerical experiments on two examples are reported.