Rostering Air Traffic Controllers

Abstract

Many, if not most, real world scheduling problems fall into the class of NP. Classical, mathematically exact methods when applied to these problems often suffer from scaling issues, that prevent the computation of a solution in reasonable time. Similarly, real world problems can often be characterised by the requirement to dynamically change any prebuilt roster in reaction to unforeseen changes to the resources available, such as short notice staff absence or a change in the requirements of the task to be fulfilled. Metaheuristic methods have been successfully used to make approximations to optimal solutions which are good enough for practical use. In this paper an example of a novel and complex employee scheduling, or rostering, problem will be discussed and a simple metaheuristic method demonstrated to solve the problem.