Multi-Commodity distribution under uncertainty in disaster response phase: Model, solution method, and an empirical study

When earthquakes and other natural disasters occur, a surge in demand for life-supporting commodities can occur alongside transportation network disruption. Slow and ineffective delivery of water, food, medical supplies, and survival equipment can lead to high levels of anxiety in the population, mistrust for the government, and potentially social conflict and increased casualties. Thus, the issue of how to effectively and efficiently distribute essential commodities in the highly stochastic post-disaster response phase is extremely critical. In this research, we create a network flow model which allows the transportation of goods not only from distribution center to relief centers, but also among the relief centers. This structure allows local relief centers to support each other, which we show leads to much greater efficiency of commodity distribution to meet the demands of each relief center in the face of uncertain transportation network flow. We collaborate with the National Science and Technology Center for Disaster Reduction (NCDR) in Taiwan to model and simulate the relationship between earthquake attributes (e.g., magnitude, time of strike) and the resulting status of the transportation network and speed of vehicle traffic. By applying simulation optimization techniques to solve the split delivery multiple destination inventory routing problem, we are able to obtain the best vehicle and inventory routing decision under varying disaster scenarios. The resulting actionable insights can alleviate civil unrest and anxiety as well as save lives after a major earthquake.