A globally convergent approximately active search algorithm for solving mathematical programs with linear complementarity constraints

摘要

In this paper, we propose a new algorithm for solving mathematical programs with linear complementarity constraints. The algorithm uses a method of approximately active search and introduces the idea of acceptable descent face. The main advantage of the new algorithm is that it is globally convergent without requiring strong assumptions such as nondegeneracy or linear independence condition. Numerical results are presented to show the effectiveness of the algorithm.