A Variant Of The Primal Transportation Algorithm*

Publisher Website
Google Scholar
Abstract
Transportation problems with varying demands are solved in two steps: first the primal algorithm is used to find an initial optimal solution, and secondly the solution is reoptimized for the altered parameters by means of the dual method. However, recent computation studies clearly indicate that the dual transportation algorithm is less efficient than the primal method. It is relatively easy to interpret the post-optimization problem as a transshipment problem. Consequently, an efficient method is available in the form of the primal network code. In this paper, an alternate primal solution procedure is proposed. A simple variant of the primal transportation algorithm is given which solves both problems, the initial optimization as well as the post-optimization. This modified version is then tested within the framework of the most efficient computer codes for the primal method.