Systems and Industrial Engineering, University of Arizona
Algorithmic Challenges in Stochastic Programming
Abstract
Stochastic Programming refers to a class of constrained
optimization problems in which some data may be uncertain, and are
modeled using random variables. The deterministic equivalent of such
problems may lead to very large scale (even infinite dimensional)
problems. Successful algorithms for such problems rely on successive
approximations which provide solutions that are optimal asymptotically.
We will discuss some basic algorithms, and then discuss their
limitations. The challenges discussed in this talk will arise from
computational considerations.
