Integration of genetic algorithms with airbase simulations for repair time analysis

Simulations have been used to investigate many important Air Force logistics issues such as personnel and resource procurement and allocation, logistics planning and policy analysis. Traditional simulations are descriptive in nature, but can be used prescriptively through what-if analysis. The quality of the resulting solution is clearly dependent on the alternative configurations considered in the analysis. Manual tweaking of system parameters in what-if analysis to generate good alternatives for performance analysis is inefficient. In this paper, we present an approach that integrates genetic algorithms with discrete-event airbase simulations for exploring the solution space in the context of aircraft repair time analysis. The article describes the overall approach and the associated computer implementation and presents an evaluation.