Managing volunteer convergence at disaster relief centers

Disaster relief centers (DRCs) are the retailers of disaster relief supply chains. Also known as Points of Distribution (PODs), disaster relief centers are ad hoc distribution sites set up in shopping mall parking lots, schools, churches, or warehouses where relief supplies are distributed to beneficiaries. Perhaps the most significant challenges that DRCs face are volunteer and material convergence, which refer to the mass movement of people and supplies, respectively, into affected areas following large-scale disaster events. The handling of donated supplies at DRCs and their subsequent distribution to beneficiaries are carried out almost exclusively by volunteers who also show up and abandon DRCs at random points in time. This paper concerns the assignment of volunteers to tasks at DRCs and takes the following three sources of uncertainty into account: (i) beneficiary arrival and service times, (ii) donation arrival and service times, and (iii) volunteer arrival and abandonment times. We represent the DRC as a queuing system with two parallel queues, one for donors/donations and the other for beneficiaries, and analyze the system in an agent-based simulation environment. In particular, we examine the effectiveness of heuristic policies for assigning volunteers to parallel queues in which performance metrics related to the average time donors and beneficiaries spend in the DRC are minimized. Through extensive computational experimentation, we identify the most effective assignment policies under a variety of experimental conditions.