Performance analysis in the selection of imagery intelligence satellites

In 1999, the National Reconnaissance Office was ready to buy a complete replacement for its system of photoreconnaissance satellites. A number of different proposals had been submitted by different companies. The authors helped plan and execute the performance analysis component of this multi-billion-dollar source selection decision. Our framework was a multiattribute value decomposition. The metrics were evaluated using a wide variety of techniques, including linear, non-linear, and integer optimization, continuous and discrete-event simulation, network flow analysis, and a genetic algorithm. This analysis incorporated the diverse needs of intelligence users to ensure an affordable system meeting the critical needs.