Search

Among the many tools of the operations researcher is the transportation algorithm which has been used to solve a variety of problems ranging from shipping plans to plant location. An important variation of the basic transportation problem is the transportation problem with stochastic demand or stochastic supply. This paper presents a simple approximation technique which may be used as a starting solution for algorithms that determine exact solutions. The paper indicates that the approximation technique offered here is superior to a starting solution obtained by substituting expected demand for the random variables.

A linear approximation model is developed for transportation problems with stochastic demand where integer solutions are required. The technique reduces the stochastic integer programming problem to a deterministic linear approximating problem which can be solved as a capacitated transportation problem. Either the transportation algorithm or the primal-dual algorithm may be used thereby insuring integer solutions.

An approach is developed for determining tight upper and lower bounds on the stochastic variables in transportation problems with stochastic demand and penalty costs. Both a graphical analysis and iterative procedure are used in conjunction with a numerical example to illustrate the approach. This paper extends the results of an earlier paper by the author.