An integer programming approach to long range shipbuilding scheduling.

Abstract

This thesis presents an integer programming model to help the Navy develop long-range shipbuilding plans. The model is of a general nature, but is proposed specifically as a decision aid for the developers of the Navy's Extended Planning Annex (EPA). The EPA sets forth planned ship purchases five to 20 years in the future. It is currently produced with a mainly manual process that takes weeks at a time, hence it is extremely difficult for the EPA planners to respond quickly to changes in the given data and assumptions. The optimization model suggests delivery dates for new ships, based on given budgets and requirements, and accounts for such complexities as the extra costs of building a leadship or of resuming construction after a production break. The model has been formulated with the General Algebraic Modeling System (GAMS) and effectively solved with two commercial optimization packages. It performs fact enough to allow the planner to make several "what if" runs in the course of developing the EPA.