New SRQP Method for Solving Nonlinear Complementarity Problems

Source abstract

Extending the framework developed in [5], this study introduces a refined variant of the square root quadratic proximal (SRQP) method for the resolution of nonlinear complementarity problems. Unlike the conjugate gradient scheme adopted in [5], which constructs search directions through a combination of current and preceding iterates, the proposed approach incorporates a simplified correction mechanism that avoids such combinations. Under the assumption that the mapping $F$ is pseudomonotone, the global convergence of the method is rigorously established. Furthermore, preliminary numerical experiments are reported to assess the efficiency and robustness of the proposed algorithm.