The Quota Allocation Model: The Linear Optimization of a Markov Decision Process

One of the many needs of the Air Force is advanced technical degrees. These degrees can be acquired in three ways: the Air Force can directly recruit personnel with the required degrees; Air Force personnel can obtain them during off duty time from local civilian colleges near their base; or the Air Force can provide advanced academic degrees (AADs) through the Air Force Institute of Technology (AFIT) or AFIT-sponsored programs. In 1995, the AFIT Commandant initiated a re-engineering study to review the AFIT mission. One of the initiatives of that study was the Quota Allocation Model (QuAM). The QuAM model is a two-phase mathematical model based on a Markov process that is used to feed a linear optimization. Outputs from the model provide the minimum number of officers, by grade and academic specialty, that must be educated annually to meet the needs and requirements of the Air Force in each of the Air Force education codes. This thesis effort entails: developing a user-friendly tool; migrating the model from lines of FORTRAN 77 code to an Excel spreadsheet environment; highlighting the assumptions necessitated by the Markov decision process; and testing for sensitivity to variations in model input parameters (AAD requirements, attrition, and inventory factors)