Solving train scheduling problem by using simulation optimization

Train scheduling is an important part of railway management which aims to determine a feasible timetable for a set of trains while considering track capacities and operational constraints. Train scheduling is a highly complex problem due to reasons such as a large number of variables and constraints, a vast solution space, conflicting objectives, limited resources, safety protocols, and the constant need for maintenance. In this thesis, the existing literature on train scheduling problems is reviewed and the approaches used to tackle these challenges are examined. To address these challenges, a simulation optimization framework is considered. This approach allows us to account for the inherent stochasticity of railway systems. In this thesis, the aim is to build upon the proposed problem and simulation model by Yalçınkaya (2010) to improve the results and identify the most suitable metaheuristic algorithm. A comprehensive analysis of various metaheuristic algorithms is conducted, including Genetic Algorithm (GA), Simulated Annealing (SA), and Tabu Search (TS), to determine their effectiveness in solving train scheduling problems. Hybridization strategies are also explored, by combining these algorithms to harness their respective strengths and improve solution quality. The findings demonstrate that integrating simulation with metaheuristics offers a promising avenue for optimizing train schedules, resulting in more efficient rail systems. In conclusion, this thesis sheds light on the critical importance of efficient train scheduling in contemporary rail systems and presents an approach that combines simulation and metaheuristic algorithms to address the inherent complexities of the train scheduling problem. The research paves the way for further investigations into optimization techniques and their practical applications in real-world rail networks.