A Study on the Feasibility of Using a Genetic Algorithm to Generate Realistic Air Traffic Scenarios Based on Recorded Field Data

Abstract

traffic scenarios based on recorded live data are used for the development and analysis of decision support tools used by air traffic controllers. Frequently these scenarios need to be modified in order to create aircraft-toaircraft encounters that are not present in the live data. This paper shows that a genetic algorithm can be used to time shift the flights in a scenario in order to create encounters with specific constrained characteristics. For this study these constraints were the distributions of the closest point of approach and the encounter angle of the encounters. The paper first describes how a genetic algorithm was implemented and then presents the results of two series of tests. The first series of tests were designed to determine if the implementation of a genetic algorithm could solve the problem. The second series of tests were designed to assess the time it took the implementation to solve the problem. The results of the study showed that a genetic algorithm could solve the problem in a reasonable time.

Original document

The different versions of the original document can be found in:

• http://pdfs.semanticscholar.org/d0f8/89be134cf7d94215da6f51497c057ce3a49b.pdf

• http://arc.aiaa.org/doi/pdf/10.2514/6.2002-4767,

http://acy.tc.faa.gov/cpat/docs/aiaa_4767.pdf,

https://academic.microsoft.com/#/detail/2163437877

• http://arc.aiaa.org/doi/pdf/10.2514/6.2002-4767,

http://dx.doi.org/10.2514/6.2002-4767