A Stochastic Service Network Design Model for Disaster Logistics Planning: A Case Study for the State of South Carolina

In this talk, we present a logistics planning problem focused on prepositioning essential relief commodities in anticipation of a hurricane landfall. We model the problem as a service network design model under the framework of two-stage stochastic programming with recourse. In the first stage, the model works to optimize the prepositioning of relief commodities for all periods, and in the second stage, it focuses on demand fulfillment, demand shortage levels, and transportation flows decisions. We assume that the hurricane’s evolution over time can be approximated as a Markov chain, where each Markovian state is characterized by the hurricane’s attributes (location and intensity). Demand quantities at each point are calculated based on these evolving attributes, allowing for more accurate scenario generation. To solve the model efficiently, we apply Benders decomposition for improved computational performance. Additionally, we implement a rolling horizon approach for adaptive decision-making as the hurricane’s forecasted path and intensity are updated over time. Our numerical results and sensitivity analyses based on a case study for the state of South Carolina demonstrate the effectiveness of this adaptive, scenario-based approach compared to a deterministic service network design model.