Heuristic Search for Tactical Runway Configuration Management

Abstract

Tactical Runway Con guration Management (TRCM) plans runway con guration (groups of runways) usage over a pre-speci ed time interval (less than ve hours) to minimize arrival and departure delays while taking into account constraints including ight patterns, taxi plans, aircraft, weather, and airport usage. Previous e orts focused on an exhaustive recursive search to test each runway con guration and determine the best possible management scheme. Here, a collection of heuristic search procedures are compared to a recursive search with respect to solution quality and computational time. The speed with which a high quality management scheme is found for a metroplex (a collection of airports) is critical. The TRCM must be evaluated several times during the day and the optimization routines must run e ciently in order to provide timely information to air tra c managers. An additional di culty is that TRCM decision procedures are not uniform across airports. However, the fundamental goal remains the same for every airport: to select an airport conguration to maximize overall e ciency of the runways, airport surfaces, terminal airspace, and interaction of the airport with the National Airspace System. Di erences between airports arise in the way runway con guration decisions are made, airport surfaces are used, and terminal airspace is managed changes between airports. Despite these di erences, a Tactical Runway Con guration Management optimization routine must be applicable to any airport (or metroplex) with limited adjustments. Test cases based on data from John F. Kennedy International Airport in New York City are provided. We report the bene ts of using heuristic search procedures over an exhaustive recursive search.