4:25 PM / Salon A/B

Tuesday, May 21

## IP4

Stochastic Programming - A Permanent Challenge

Stochastic programming is an area which, on the one hand, is closely linked to applications, and on the other hand, serves as a source of challenging theoretical and computational problems.
Many practical decision problems have uncertain parameters; when they are modeled as random variables, stochastic programming problems arise. Various ways of formulating the objective and the constraints and different information patterns lead to a wealth of problem formulations. Some of them involve infinite dimensional spaces and nonsmooth or discontinuous functions, which inspires new theoretical developments.

Solutions to stochastic programs can hardly be obtained by standard linear or nolinear programming techniques; due to the large size and/or involved structure specialized methods are necessary. The area of stochastic programming is a laboratory for developing new decomposition methods for large scale problems and new techniques that combine simulation and optimization.

**Andrzej Ruszczynski**

International Institute for Applied Systems Analysis

Austria

*MEM, 3/11/96*