Stochastic Optimization Problems with Partially Known Distribution Functions
Authors: Ermoliev YM, Nedeva C
Publication Year: 1982
Reference: IIASA Collaborative Paper CP-82-060
Abstract
The main purpose of this paper is to discuss the numerical procedures of optimization with respect to unknown distribution functions which are based on ideas of duality. Dual problem is formulated as minimax type problem without concavity of "inner" problem of maximization. Numerical procedure allowed to avoid difficulties concerning "inner" problem solution is proposed.