Planning Under Uncertainty: Solving Large Scale Stochastic Linear Programs

Efficient technologies have been developed to solve large deterministic dynamic linear systems, but deterministic methodologies cannot account for real uncertainties. Stochastic models avoid this shortcoming but are often computationally expensive or intractable. An approach devised by George Dantzig, P. Glynn, and Gerd Infanger and explained in this book is capable of solving large scale stochastic linear problems with numerous stochastic parameters. The method hs been successfully tested on facility expansions and financial planning problems. In one case a problem whose deterministic equivalant would appear as a linear programming problem with approxiamately 10/27 constraints and variables was solved on a laptop 80386 computer.