THE CONVERGENCE OF APPROACH PENALTY FUNCTION METHOD FOR APPROXIMATE BILEVEL PROGRAMMING PROBLEM

In this paper, a new algorithm-approximate penalty function method is de- signed, which can be used to solve a bilevel optimization problem with linear constrained function. In this kind of bilevel optimization problem, the evaluation of the objective function is very difficu1t, so that only their approximate values can be obtained. This algorithm is obtained by combining penalty function method and approximation in bilevel programming. The presented algorithm is completely different from existing methods. That convergence for this algorithm is proved.