Simultaneous location and vehicle fleet sizing of relief goods distribution centers and vehicle routing for post-disaster logistics

• Published in: Computers & operations research 2024-01, Vol.161, p.106404, Article 106404
• Main Authors: Chang, Kuo-Hao, Chiang, Yi-Chieh, Chang, Tzu-Yin
• Format: Article
• Language: English
• Subjects: Disaster operations management; Humanitarian logistics; OR in disaster relief; Sample average approximation; Simulation optimization
• ISSN: 0305-0548
• DOI: 10.1016/j.cor.2023.106404

The occurrence of a sudden-onset disaster such as a major earthquake brings about a great deal of uncertainty in, for example, the severity of damage to the road network and other infrastructure across the disaster region, the traffic flow conditions in the post-disaster network, and the demand for relief goods at relief centers. Despite these uncertainties, crucial decisions must still be made both before and rapidly after the occurrence of the disaster such that the logistical operation of relief goods distribution can have the greatest possible chance of success. In this research, we propose a two-stage stochastic programming model in which the first stage optimizes the location of relief goods distribution centers as well as the number of vehicles allocated to the logistical operation, while the second stage determines the best vehicle and inventory routing decisions for the logistics operation in the critical first time window after the random factors associated with the disaster are revealed. We collaborate with the National Science and Technology Center for Disaster Reduction (NCDR) in Taiwan to effectively model post-disaster road network conditions and the speed of vehicle traffic as a function of earthquake parameters (e.g., magnitude, time of strike). An efficient simulation optimization algorithm with feedback is proposed to tackle the two-stage stochastic programming model. We carry out an empirical study that consists of a set of experiments based on real data from NCDR to investigate the impact of critical factors in the proposed model under a variety of different conditions. Managerial insights are derived for decision makers pertinent to both the preparation and response phases of a sudden-onset disaster. •A two-stage stochastic programming model was proposed for disaster operations management.•A simulation optimization solution framework was developed.•An empirical study was conducted in collaboration with NCDR in Taiwan.•Useful managerial insights were gained for decision makers.