Multivariate Manpower, Personnel and Training (MPT) Modeling and Managment System

The challenge is to substantially increase the Navy's efficiency and effectiveness in personnel management. IET will develop decision-theoretic models that trade-off the requirements of current and future military assignments against the capabilities and needs of military personnel under conditions set by Navy policy. These decision-theoretic models will combine Bayesian network technology, to model the uncertainties of the situation, with Multi-Attribute Utility, to model the priorities in the trade-offs. The advantages of IET's approach are: (1) a flexible yet comprehensive modeling system capable of probabilistic computation over a wide range of policy, requirement and service member variables, (2) the ability to predict the trainability, performance, and reenlistment likelihood of service members, (3) the use of Internet system interface technology to maximize the reuse of legacy systems and to provide personnel management modeling capability anytime and anywhere an Internet browser interface exists, (4) the ability to evolve models over time, and (5) the ability to learn model parameters from data as well as experts. During Phase I, IET will develop models that predict trainability, performance and re-enlistment likelihood. These will become the foundation for more complex models in Phase II that will not only scale, but will also evolve as requirements change. The primary application and organizational benefit of this research will be for agencies and organizations (both Government and Commercial) with responsibility for significant long-term management of personnel.