不確定下での配分計画の一解法

Abstract

The distribution problem whose constsained resource vector is random variable can be solved by using an algorithm, developed by Dantzig and Madansky, called "tow-stage programs under uncertainty conditions". It's certainty equivalence problem, however, would be formulated into nonlinear programming problem and it is difficult to treat for it's nonlinearity. In this paper, we propose an algorithm to solve the distribution problem under uncertainty by considering the characteristic of the objective function of the problem. This algorithm consists of iterations of solving a simple simplex method and is more advantage than general procedure solving non-linear programming in view of both the computing time and the computer storage required.