Optimization of Functionals Which Depend on Distribution Functions: 1. Nonlinear Functional and Linear Constraints

The main purpose of this paper is to discuss numerical optimization procedures for problems in which both the objective function and the constraints depend on distribution functions. The objective function is assumed to be nonlinear and to have directional derivatives, while the constraints are taken as linear. The proposed algorithm involves linearization of the objective function at the current point and solution of an auxiliary linear subproblem. This last problem is solved using duality relations and cutting-plane techniques.