Designing a humanitarian supply chain network with integrated inventory and transportation resilience

Natural and man-made disasters occur every year worldwide and cause significant damage to affected regions and populations. Providing timely humanitarian assistance after disasters requires an efficient humanitarian relief supply chain network. However, the unstable conditions of disaster-affected environments may disrupt relief operations and delay the delivery of required aid. Therefore, developing appropriate resilience strategies is essential to reduce the effects of such disruptions. This study presents a resilient humanitarian supply chain network design model that determines the optimal locations of production and distribution centers and addresses the allocation of raw materials and relief items among suppliers, producers, distribution centers, and affected regions. Two resilience strategies are considered in the proposed framework, including maintaining inventory at distribution centers and selecting reliable transportation routes across different levels of the network. The problem is formulated as a mixed-integer linear programming model. In addition, the Analytic Hierarchy Process is employed to identify reliable transportation routes. Initially, the proposed model is validated through a numerical example solved by the CPLEX solver. After that, three metaheuristic algorithms, including genetic algorithm, differential evolution, and particle swarm optimization, are designed using a problem-specific solution representation scheme to solve larger instances of the problem. The computational results confirm the robustness of the -based route selection, where the main route rankings remain relatively stable under changes in criterion weights and the reliability threshold. Moreover, the results show that the genetic algorithm has the best overall performance among the proposed metaheuristics. This algorithm reaches the CPLEX optimal solution in 60% of the small-sized instances and obtains the lowest mean relative percentage deviation for medium- and large-sized problems while maintaining competitive computational times.