Practical train crew scheduling using improved tabu search

This paper proposes improved tabu search for practical train crew scheduling problems (CSPs). The proposed method includes a new objective function, a new generation method of an initial schedule, and new generation methods of neighborhood schedules to apply to practical problems. The results indicate that the proposed method can generate an appropriate schedule which satisfies all of constraints, and drastically reduce the objective function values for an actual timetable than the conventional method.