Modeling and solution algorithm for a disaster management problem based on Benders decomposition

Pre-disaster planning and management activities may have significant effects on reducing post-disaster damages. In this article, a two-stage stochastic programming model is provided to design a resilient rescue network assuming that the demands for relief items and the network functionality after the disaster are affected by uncertainty. Locations and capacities of relief centers, the inventory of relief items, and strengthening vulnerable arcs of the network are among the main decisions that must be taken before the disaster. Servicing the affected points is decided after the disaster, and the risk of not satisfying demands is controlled by using the conditional-value-at-risk measure. Since the direct resolution of the model is intractable and time-consuming over actual large-sized instances, an improved Benders decomposition algorithm based on the problem structure is proposed to overcome this difficulty. Computational results highlight the effectiveness of the proposed method compared to the existing approaches.