A new computational method for the functional inequality constrained minimax optimization problem

In this paper, the authors consider a general class of functional inequality constrained minimax optimization problems. This problem is first converted into a semi-infinite programming problem. Then, an auxiliary cost function is constructed based on a positive saturated function. The smallest zero of this auxiliary cost function is equal to the minimal cost of the semi-infinite programming problem. However, this auxiliary cost function is nonsmooth. Thus, a smoothing function is introduced. Then, an efficient computational procedure is developed to estimate the smallest zero of this auxiliary cost function. Furthermore, an error bound is obtained to validate the accuracy of the approximate solution. For illustration, two numerical examples are solved using the proposed approach.