Train Scheduling with the Application of Simulated Annealing

In this paper, we are going to schedule the train movement in one-track routes, in which, the trains could arrive to their destination at the earliest and safest way.In this problem, we consider two different directions for the trains. In the other word, there are southbound trains which travel form north to south and also northbound trains that travel from south to north. Furthermore the origins and destinations of all these trains could be different but they should be one of the stations defined in the problem.The train-scheduling problem that is defined in this project is an integer-programming problem that could be solved by current solving methods such as Brach & Bound.As the required solving time of above mentioned methods which find the optimum answer increases exponentially by the size of the problem, in order to find a good answer in an acceptable time for the real problems, Meta Heuristic algorithms such as Simulated Annealing are useful. In this paper after presenting the explanation and mathematical modeling of the train-scheduling problem, the application of the simulated annealing algorithm for problem is presented. In the next section, after tuning the parameters of the proposed algorithm using Design of Experiments method, we compare the final result of the proposed algorithm with the upper bound produced by branch and bound method using Lingo 8.0 Software. To illustrate the efficiency of the exhibited algorithm, Tehran/Bandar-abbas rail route has been selected to be scheduled. Finally the resulted timetable shows a wide improvement in comparison with the applied timetable prepared by the trail and error method used by the RAI.