Date of Award

Spring 1-1-2015

Document Type


Degree Name

Doctor of Philosophy (PhD)


Operations & Information Management

First Advisor

Thomas W. Vossen

Second Advisor

Manuel Laguna

Third Advisor

Stephen R. Lawrence

Fourth Advisor

Dan Zhang

Fifth Advisor

Marco Better


Civilian and military manpower planning is commonly conducted by human resource (HR) planners having little or no experience using and developing quantitative methods. If an organization forecasts an increase or decrease in production, HR planners must respond with recommended changes to the employee footprint. It is vital for planners to properly identify personnel needs by job skill and tenure in order to properly recommend the required changes in total personnel. The use of mathematical models can provide valuable support when making manpower planning decisions, such as personnel recommendations based on characteristics of a set of employees or issuance of incentive bonuses based on a company's skill shortages.

In the manpower planning process two important levers are available to a company to shape its workforce-recruitment and retention of personnel, and these can be employed individually or collectively. For example, a company may augment its workforce or replace departing personnel by recruiting new employees, or it can seek to retain incumbent employees by incentivizing or de-incentivizing them with respect to their retention decisions. In this dissertation, we first present mathematical models that can assist civilian and military personnel planners in balancing recruitment and retention. Both hiring and retention incur costs: monetary compensation may be required to entice current employees to remain, and recruiting employees will almost certainly involve occupational training and reduce productivity until this training is complete. Thus, in both military and civilian sectors, this choice as to whether to hire or retain can be of the utmost importance.

Effective manpower planning requires a thorough understanding of employee retention behavior, that is, the likelihood that employees will stay with the organization. In this dissertation, we therefore also propose models that can assist HR planners in estimating retention probabilities of employee-groups fitting particular profiles. Accurately forecasting the retention propensity of a group of employees is key because it helps a company prepare for the possible departure of these employees, while also identifying possible reasons (via a set of attributes) motivating their decision to leave.

The issues outlined above may require the use of large data sets, which can complicate real- time analysis efforts. As such, the final chapter of this dissertation presents methods for extracting a representative subset of employee data. The resulting data sets can simplify subsequent analysis efforts and enable real-time analysis of retention and recruitment decisions. The resulting models have applications beyond manpower planning and may provide a general framework for alleviating the burden of conducting analysis using large data sets in a variety of settings.