An optimal allocation of Army recruiting stations with active and reserve recruiters

Abstract

This thesis addresses the problem of how to locate and staff recruiting stations with Active and Reserve recruiters in order to maximize the annual number of recruits. The problem is formulated as a nonlinear integer programming problem. The objective function for the problem, also referred to as the production function, describes the number of recruits obtainable from each zip code and can be estimated via Poisson regression. The resulting nonlinear integer programming problem is heuristically solved by decomposing decision variables into two sets: one to locate stations and the other to staff them with recruiters. Comparisons are made between problems with production functions derived from all zip codes and those derived from only zip codes belonging to efficient stations as defined in Data Envelopment Analysis.