Stochastic network optimization for strategic resource pre-positioning and allocation

Abstract

This paper presents a stochastic network modeling approach to develop insights into strategic facility location planning, capacity management, resource pre-positioning, and allocation. The primary purpose of the proposed model is to present a cost-effective logistics network designed for efficiently handling diverse relief items across a spectrum of crisis scenarios. By incorporating stochastic elements, we aim to capture the inherent unpredictability of demand fluctuations and the impact of crises. Our approach optimizes facility sizes to leverage economies of scale while improving allocation decisions. Additionally, it ensures fairness across demand points by implementing a strategy to mitigate relative shortages. To demonstrate the practical applicability of our model, we conduct a computational case study utilizing instances from the national food stockpiling system in Germany. Moreover, we present a sensitivity analysis highlighting the impact of crisis intensity, increased storage and production capacity, and weighting decisions of transportation costs on facility location and assignment decisions. The results provide economic and managerial insights for public decision-makers, enhancing cost-effective disaster preparedness and network design. The case study shows that the proposed model optimizes inventory by eliminating excess quantities and favoring large warehouses, reducing costs through fewer locations. However, prioritizing rapid delivery results in a more decentralized network with smaller, costlier warehouses. The logistics network adapts to varying demand scenarios, strategically placing warehouses in densely populated regions with higher crisis risks.