Inventory relocation for overlapping disaster settings in humanitarian operations

The number and scale of humanitarian operations has significantly increased during the past decades due to the rising number of humanitarian emergencies and natural disasters worldwide. Therefore, the development of appropriate planning methods for optimization of the respective supply chains is constantly growing in importance. A specific problem in the context of humanitarian operations is the supply of relief items to the affected areas after the occurrence of a sudden change in demand or supply, for example, due to an epidemic or to unexpected shortages, during an ongoing humanitarian action. When such overlapping disasters occur, goods must be relocated to existing depots in a way which enables rapid supply to regions with new and urgent demand. At the same time, ongoing operations have to continue, i.e., the other regions should not suffer from shortages, and possible future emergencies must be taken into account. This is a planning situation under uncertainty as it is not known in advance if and where a disruption-and hence additional demand-will occur. In this paper, an optimization model for such situations is developed based on penalty costs for non-satisfied demand. A rolling horizon approach for solving the model is presented, and it is shown that taking into account the possibility of future disruptions can help to balance inventories and to reduce total non-served demand.