Combining Generalized Programming and Sampling Techniques for Stochastic Programs with Recourse

Gaivoronski, A.A. & Nazareth, J.L. (1986). Combining Generalized Programming and Sampling Techniques for Stochastic Programs with Recourse. IIASA Working Paper. IIASA, Laxenburg, Austria: WP-86-044

Preview

Text WP-86-044.pdf Download (527kB) | Preview

Abstract

This paper deals with an application of generalized linear programming techniques for stochastic programming problems, particularly to stochastic programming problems with recourse. The major points which needed a clarification here were the possibility to use the estimates of the objective function instead of the exact values and to use the approximate solutions of the dual subproblem instead of the exact ones.

In this paper conditions are presented which allow to use estimates and approximate solutions and still maintain convergence. The paper is a part of the effort on the development of stochastic optimization techniques at the Adaptation and Optimization Project of the System and Decision Sciences Program.

Actions (login required)

View Item