On solving optimization problems with hidden nonconvex structures

Abstract

Here we consider three very popular optimization problems: the linear complementarity problem, the search for Nash equilibria in a bimatrix game, and the quadratic-linear bilevel programming problem. It can be shown that each of the problem possesses a hidden nonconvexity and, as a consequence, a rather large number of local solutions which are different from global ones from the viewpoint of the goal function. In order to attack these problems the principal points of Global Search theory are presented and discussed. Furthermore, the main stages of Local and Global Search Methods are precised for each problem. Finally, we present the new results of computational solution separately for every problem considered.