Reverse Engineering of Foreign Missiles VIA Genetic Algorithm

摘要

One mission of the National Air Intelligence Center (NAIC) is the reverse engineering of foreign missile weapon systems from incomplete observational data. In the past, intuition and repeated runs of a missile performance model were required to converge to a solution compatible with observed flight characteristics. This approach can be cumbersome and time consuming, as well as being subject to undesirable influences from the analyst's preconceptions and biases. An alternative approach has been created to apply genetic algorithm (GA) techniques to allow automation of the process, wider exploration of the design space, and more optimal solutions matching the observational data. The GA, when interfaced with a missile performance model, was able to identify a set of missiles that very closely matched the observed performance of a given sample missile. The approach was able to provide the analyst with multiple candidate missiles for further analysis that would have been missed by the previous trial and error approach.