Value of sorting and recovery in post-disaster relief aid distribution

In-kind donations gathered from the public after a disaster typically consist of an unknown and disorganized composition of both "useful"and "useless"items. Sending the donations directly to the affected people without any pre-processing leads to not only extra shipment and handling costs but also congestion in the disaster areas, and even chaos in local distribution. Motivated by this fact, we investigate the effect of relief aid sorting and subsequent recovery of items that do not fulfill immediate needs under demand, donation quantity, and donation type uncertainties. To this end, we propose a supply network configuration with sorting centers for pre-processing and optimize it by a three-stage stochastic programming model. We compare the case of the sorted donations with the case where donations are directly sent to points of distribution to victims. We apply the proposed model to the Istanbul earthquake case under different parameter settings and derive various managerial insights. We observe that sorting the donated relief aid items and sending only the useful ones to disaster victims may decrease the expected total cost, in addition to some intangible benefits, but at the same time may increase the expected time of the deliveries. However, it is possible to cope with this response time deterioration by increasing the workforce level at the sorting centers with an acceptable additional cost.