NETWORK OPTIMIZATION TO MODEL RANDOM RISK OF SUPPLY CHAIN DISRUPTIONS

Abstract

The U.S. Navy’s supply chain stretches globally, supporting the fleet in multiple theaters to enable sustained forward presence, security, and deterrence. However, supply chains are subject to disruptions that slow materiel movements throughout the network, and these disruptions may severely hinder the readiness of ships operating in distant theaters. A common culprit for peacetime supply chain disruptions is adverse weather, which is especially true in waters that are prone to major tropical storm systems. Other disruptions may include failure of equipment, accidents, and adversarial activity during active conflict situations. With these concerns in mind, this thesis formulates six optimization models to assist logistics planners in preparing for and responding to these uncertain contingencies. The models we present fall into both a proactive family, which plan for disruptions based on their likelihood before they occur, and a reactive family, which respond to the disruptions as they occur. To address the probabilistic risks of disruptions, these models utilize linear integer programming, chance constraints programming, and dynamic programming in different ways, seeking to demonstrate various methods for routing supplies through a network vulnerable to random disruptions. Lastly, we analyze results to determine the suitability of these models in several disruption scenarios.