A NEW KERNEL FUNCTION GENERATING THE BEST COMPLEXITY ANALYSIS FOR MONOTONE SDLCP

In this article, we propose a new class of search directions based on kernel function to solve the monotone semidefinite linear complementarity problem by primal-dual interior point algorithm. We show that algorithm benefits from best polynomial complexity, namely $ O(\sqrt{n}(\log n)^{2}\log\dfrac{n}{\epsilon})$. The implementation showed great improvement concerning time and number iterations.