Air Traffic Controller Shift Scheduling by Reduction to CSP, SAT and SAT-Related Problems

摘要

In this paper we present our experience in solving Air Traffic Controller Shift Scheduling Problem. We give a formal definition of this optimization problem and introduce three encodings. The encodings make possible to formulate a very wide set of different scheduling requirements. The problem is solved by using SAT, MaxSAT, PB, SMT, CSP and ILP solvers. In combination with these solvers, three different optimization techniques are presented, a basic technique and its two modifications. The modifications use local search to modify some parts of the initial solution. Results indicate that SAT-related approaches outperform other solving methods used and that one of the introduced techniques which uses local search can significantly outperform the basic technique. We have successfully used these approaches to make shift schedules for one air traffic control center.