Tuesday, May 21
8:00-10:00 AM
Salon B
MS17
Approximation and Computation in Stochastic Programming
(Invited Minisymposium)
Stochastic programming is a useful tool for planning under uncertainty. Applications may be
found in finance, energy systems, transportation, capital expansion, inventory management,
scheduling, to name a few. The mathematical technology is based on large scale optimization,
linear or quadratic programming, with special techniques for exploiting problem structure and
parallelism.
The basic difficulties in the field concern the approximation of the probability distribution
and the solution of the resulting large scale problem. Applications have been slow to develop
due to the complexity of the problem the method is addressing (multiple stage decisions under
uncertainty). But in the fields most accustomed to coping with risk and uncertainty, like
finance, there is a growing conviction that this is the only technology that can address all
the issues.
The speakers in the symposium will discuss issues of approximation and computation unique to
the field.
Organizer: Alan J. King
IBM T.J. Watson Research Center
- Stochastic Programming Models in Practice
- John R. Birge, University of Michigan
- Epigraphical Limit Laws in Stochatic Programming
- Lisa Dorf, University of California, Davis
- Density Estimation: A Bayesian Approach
- Michael X. Dong, University of California, Davis
- Simulating Empirical Distributions for Multistage Stochastic Programs
- Alan J. King, Organizer
LMH, 3/15/96