Mixed Integer Linear and Nonlinear Optimization for Disadvantaged Populations with Accents of Fairness and Balance
by Shima Azizi
The use of mixed-integer optimization to help disadvantaged populations is a relatively new area of study in the operations research domain. We focus on the use of mixed-integer linear and nonlinear optimization to take a broader view of capitalism, beyond only maximizing profits, to assist disadvantaged populations and make positive societal impacts. To this end, we investigate resource allocation, team orienteering, inventory management, and network flows to help three distinct groups of disadvantaged populations: underserved patients relying on the emergency department for their non-urgent medical needs, refugees, and foster children. First, we consider an application of team orienteering problem in a recent healthcare delivery innovation. We investigate the management of a community paramedicine program with respect to efficiency by establishing the first optimization-based framework that incorporates fairness considerations with the goals of increasing patient welfare, lowering hospital costs, reducing readmissions and emergency department visits. Second, we consider an application of humanitarian aid inventory management. We seek to allocate supply aid to refugee camps with uncertain demands and replenishment cycles 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. Third, we study a time-space network application in a humanitarian scheduling problem. We introduce a new modeling approach to schedule regular meetings between foster children and their biological parents to improve the quality of life for foster children and help non-profit foster care organizations in managing their resources efficiently.
Dr. Sharon Johnson, WPI
Dr. Renata Konrad, WPI
Dr. Erhun Kundakcioglu, Özyeğin University
Dr. Andrew Trapp, WPI (Advisor)