This paper describes the redesign of the Federal Aviation Administration's implementation of a genetic algorithm used for time shifting flights in air traffic scenarios. Time shifted scenarios are used in testing decision support tools that predict the potential loss of separation between aircraft. This paper describes the improvements that resulted when this application was redesigned and coded in Java. The improvements described in this paper include the following: • Maintainability improved as a result of a modular design using object-oriented techniques. • Usability improved as a result of more efficient logging techniques, configuration methods, and user interfaces. • Quality of the solution improved as a result of a more accurate method for calculating of aircraft-to-aircraft conflicts. • Timeliness for obtaining a solution improved as a result of using modern software engineering techniques, such as distributing the fitness function across multiple processors and caching fitness scores.
Jaguar: Time shifting air traffic scenarios using a genetic algorithm
2012-10-01
538661 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Kraftfahrwesen | 2005
|