Abstract:

A new method to  absolute value equations was presented. Firstly, absolute value equations were transformed into linear complementarity problems. Combining Newton direction and centering direction, we obtained search direction by solving a linear system, then established a feasible interior point algorithm for absolute value equations. We proved that by this method an optimal solution can be obtained after a finite number of iterations. At last, we gave some numerical examples to indicate that the method is feasible and effective.

Key words: absolute value equations, linear complementarity problem, feasible interior point algorithm, polynomial complexity