A GAMS-based model of the U.S. Army Wartime Ammunition Distribution System for the Corps level

Abstract

The U.S. Army Wartime Ammunition Distribution System (WADS) will experience an unprecedented demand for ammunition under the operational concept of Airland Battle. To meet demand, proper storage facility location and an efficient flow through the distribution network will be required. Using information from Army Field Manuals, maps and simulation data for demand, both a mixed integer program (MIP) and a sequential, optimization-based heuristic are developed to model the WADS. The Generalized Algebraic Modelling System is used to implement both models. The sequential heuristic locates ammunition facilities with a binary integer program and then directs ammunition through those facilities utilizing a network flow model with side constraints. The MIP integrates location and flow decisions in the same model. For a general scenario, the sequential heuristic locates a 21 node, 30 arc network with ammunition flows over 30 time periods in 22 CPU seconds on an IBM 3033AP. For the same scenario the MIP obtains a solution for only a 3 time period problem in 87 CPU seconds. Keywords: Ammunition, Integer programming, Heuristic, Networks