Optimizing US Army Force Size Under Uncertainty Through Stochastic Programming

摘要

We have demonstrated that multistage stochastic programming is an appropriate and interesting approach to force size analysis, providing a stochastic treatment of deployment demand uncertainty and the optimization of force size and composition over time. We developed the methodology to translate simulated deployment demand into a discrete probability distribution amenable to stochastic programming. We constructed the AFSP model, a LP model of total Army manpower requirements, implementing it with stochastic deployment demand. We derived policy constraints to accurately model the FYDP decision making cycle.