Search for: All records

There are 26 million refugees worldwide seeking safety from persecution, violence, conflict, and human rights violations. Camp-based refugees are those that seek shelter in refugee camps, whereas urban refugees inhabit nearby, surrounding populations. The systems that supply aid to refugee camps may suffer from ineffective distribution due to challenges in administration, demand uncertainty and volatility in funding. Aid allocation should be carried out in a manner that properly balances the need of ensuring sufficient aid for camp-based refugees, with the ability to share excess inventory, when available, with urban refugees that at times seek nearby camp-based aid. We develop an inventory management policy to govern a camp’s sharing of aid with urban refugee populations in the midst of uncertainties related to camp-based and urban demands, and replenishment cycles due to funding issues. We use the policy to construct costs associated with: i) referring urban populations elsewhere, ii) depriving camp-based refugee populations, and iii) holding excess inventory in the refugee camp system. We then seek to allocate aid in a manner that minimizes the expected overall cost to the system. We propose two approaches to solve the resulting optimization problem, and conduct computational experiments on a real-world case study as well as on synthetic data. Our results are complemented by an extensive simulation study that reveals broad support for our optimal thresholds and allocations to generalize across varied key parameters and distributions. We conclude by presenting related discussions that reveal key managerial insights into humanitarian aid allocation under uncertainty. 

Every year, tens of thousands of refugees are resettled to dozens of host countries. Although there is growing evidence that the initial placement of refugee families profoundly affects their lifetime outcomes, there have been few attempts to optimize resettlement decisions. We integrate machine learning and integer optimization into an innovative software tool, Annie™ Matching and Outcome Optimization for Refugee Empowerment (Annie™ Moore), that assists a U.S. resettlement agency with matching refugees to their initial placements. Our software suggests optimal placements while giving substantial autonomy to the resettlement staff to fine-tune recommended matches, thereby streamlining their resettlement operations. Initial back testing indicates that Annie™ can improve short-run employment outcomes by 22%–38%. We conclude by discussing several directions for future work. 

Employment outcomes of resettled refugees depend strongly on where they are placed inside the host country. While the United States sets refugee capacities for communities on an annual basis, refugees arrive and must be placed over the course of the year. We introduce a dynamic allocation system based on two-stage stochastic programming to improve employment outcomes. Our algorithm is able to achieve over 98 percent of the hindsight-optimal employment compared to under 90 percent of current greedy-like approaches. This dramatic improvement persists even when we incorporate a vast array of practical features of the refugee resettlement process including indivisible families, batching, and uncertainty with respect to the number of future arrivals. Our algorithm is now part of the Annie™ MOORE optimization software used by a leading American refugee resettlement agency. The full version of this paper is available at https://arxiv.org/pdf/2105.14388.pdf.