Improving short-term conflict alert via tabu search

In this paper we describe some work carried out by National Air Traffic Services in the UK into developing optimisation tools to help improve the effectiveness of one of their air traffic control safety systems – short-term conflict alert. In short-term conflict alert a computer system continually monitors radar data and alerts air traffic controllers if it detects a situation where two aircraft are in danger of approaching too close to one other. Within the computer program that makes up the short-term conflict alert system is a large number of parameters. Choosing appropriate values for these parameters is a task that is currently done via extensive human intervention. In this paper we describe how a modern heuristic technique, tabu search, can be used to make parameter choices.