The Wartime Portfolio Selection Problem




Woodaman, Ronald Frederick Alexander

Journal Title

Journal ISSN

Volume Title



This thesis describes the research conducted to support the optimal selection of a portfolio of military solutions during wartime. During peacetime, the United States military selects a portfolio of military solutions holistically as part of an annual budgetary planning cycle supported by long-term planning. During wartime, this annual review and decision process is not responsive to the rapid cycles of battlefield adaptation and the resulting exploitation of opposing capability gaps by adversaries. The long-running vulnerability of U.S. forces in Iraq and Afghanistan to Improvised Explosive Device (IED) attacks, as the threat's tactics and techniques evolved during these conflicts, provides a poignant example. During conflict, opportunities to improve the force arrive irregularly over time and are difficult to anticipate. When potential solutions are identified, these must be rapidly pursued, subject to resource constraints. Bad decisions early rob resources from better opportunities that arrive later, while good opportunities unduly delayed may lead to lost opportunities on the battlefield.



Operations research, Statistics, Computer science, Approximate dynamic programming, Dynamic stochastic knapsack problem, Multi-objective decision analysis, Sequential decision analysis, Stochastic integer programming, Stochastic optimization