Generation of realistic air traffic scenarios using a genetic algorithm

Traffic flow management decision support tools such as the user request evaluation tool (URET), developed by the MITRE Center for Advanced Aviation Systems Development, and the Center-TRACON automation system (CTAS), developed by the NASA/Ames Research Center, use simulation as a tool for development, technical assessment, and field evaluation. Air traffic scenarios, based on recorded live data, are used to test these decision support tools. Frequently the scenarios need to be modified in order to create aircraft-to-aircraft encounters and conflicts that are not present in the live data. This paper presents an implementation of a genetic algorithm that is being used to time-shift the flights within an air traffic scenario to create encounters with specific constrained characteristics. These constraints are the distributions of the horizontal and vertical closest points of approach, the encounter angle at the closest point of horizontal approach, and the vertical type of encounter. This paper describes how the genetic algorithm was implemented, including a description of the solution chromosome and of the fitness function used to measure the potential solutions. After describing the implementation, a specific example of its use is presented.