zbMATH — the first resource for mathematics

An iterative method for solving two-step discrete problems of stochastic programming with additively separable variables. (Russian. Russian original) Zbl 0742.90061
The author suggests a decomposition algorithm for the solution of discrete two stage stochastic programs in which the objective function and the constraints are separable with respect to the first stage decision variables and the second stage ones. In each iteration, the dual to the second stage problem is solved and a new feasible solution of the first stage problem is obtained as a solution of the Master program — a discrete deterministic program. The algorithm is finite if the dual programs can be solved precisely. For the opposite case, a modification based on solving the dual problems by means of stochastic approximation methods is formulated. No numerical experience is reported.

90C15 Stochastic programming
90-08 Computational methods for problems pertaining to operations research and mathematical programming
65K05 Numerical mathematical programming methods
62L20 Stochastic approximation