Skip to Main Content
The process by which U.S. military organizations select humanitarian infrastructure projects in host countries is hampered by a lack of quantitative decision-support. Addressing this problem is important because the degree of effectiveness of projects selected and completed significantly impacts geo-political implications for the U.S. and, further, directly impacts the quality of life for citizens in host countries. In this paper, a family of multi-objective optimization problems are formulated and solved to explore quantitative, system-level insights on optimal project groupings and comparisons to historical data. Objective functions are found which produce project groupings that match well with historical data, and the optimization model expresses significant information about the sensitivity of solutions in the decision-space. However, imperfection in the matching as well as insights from sensitivities suggests the need for augmentations to improve upon a single organization, optimization-only model. In particular, the use of agent-based modeling is presented for extended work to capture the unmodeled dynamics in the system.