Min/Max inventory planning for military logistics

Abstract

Military logistics is concerned with the projection and sustainment of forces in accomplishing a mission. A critical part of force sustainment is supply chain planning (SCP), which involves determining the sources of supply and defining the distribution networks that will be used to meet the sustainment requirements of the force. Although supply chain planning is well-researched in the commercial sector, the military are driven by different criteria and use different planning processes on which there is limited research. In the military domain, supply chain planning involves determining the deliveries required to meet the expected demand of a particular mission by calculating suitable delivery times and supply quantities. This process is typically done using a system called Min/Max, which calculates the deliveries using minimum (Min) and maximum (Max) inventory levels and a number of other constraints. The basic idea is to resupply the inventory of a node (location) when the stock level approaches, but doesn't fall below, the Min inventory level and order enough supplies to bring the stock level up to the Max stock holding capacity. This is traditionally a manual process that can take a group of military planners several hours or days to complete, depending on the size of the problem, and the results are susceptible to human error. This paper presents the Min/Max Planning Algorithm (MPA) which can perform automated supply chain planning using the Min/Max system. Given an SCP problem with an expected profile of demand, the MPA uses the constraints defined for each node and the expected demand to calculate a suitable order profile for that node. Each node's order profile is combined with the profiles of its dependants, providing a complete picture of the supplies that a node will require to sustain itself and its dependants over the course of the mission. These combined order profiles are then converted into a list of deliveries which provides an overall solution to the SCP problem. The MPA has been evaluated using four SCP problems of increasing complexity, which are representative of typical planning problems encountered by the Australian Defence Force. The results show that the MPA is able to solve all of these problems and keep the expected minimum stock level above a defined threshold, which provides a suitable buffer against uncertainty when the calculated deliveries are actually executed. The algorithm is also computationally efficient, being able to solve large SCP problems in less than a second. This is a significant improvement on the hours or days that it could take a military planner to perform the same calculations by hand.