Visit Allocation Problems in Multi-Service Settings: Policies and Worst-Case Bounds

Problem definition: We consider a resource allocation problem faced by health and humanitarian organizations deploying mobile outreach teams to serve marginalized communities. These teams can provide a single service or an assortment of services during each visit. Combining services is likely to increase operational efficiency but decrease the relative benefit per service per visit, as operations are no longer tailored to a single service. The aim of this study is to analyze this benefit-efficiency trade-off. Academic/practical relevance: Increased operational efficiency will enable organizations to serve more people using fewer resources. This is important given the increasing funding gap organizations are facing. Our work adds to the literature on resource allocation problems and visit allocation problems specifically, where the focus has been primarily on single services. Methodology: We analyze a general visit allocation problem incorporating demand distribution (where to go) and return time (how frequently to go). We derive analytical bounds for the benefit-efficiency trade-off, and propose visit allocation policies with worst-case optimality guarantees. Results: Our results show the benefit-efficiency trade-off can be assessed based on high level parameters. We show demand alignment is a key driver of this trade-off. We apply our results to Praesens Care, a social enterprise start-up developing mobile diagnostic laboratories, and verify our insights using real-world data. Managerial Implications: Our research contributes to the discussion on innovation and increased efficiency in health and humanitarian aid delivery by quantifying operational trade-offs in offering assortments of services. Specifically, our results help assess the potential of integrated models for health and humanitarian aid delivery and provide organizations with easy-to-implement methods to determine close-to-optimal visiting policies. Importantly, our methods remain applicable in case of limited data, making them suitable for strategic decision-making.