Microscopic Simulation of Decentralized Dispatching Strategies in Railways

This paper analyzes the effectiveness of decentralized strategies for dispatching rolling stock and train drivers in a railway system. Such strategies give operators a robust alternative in case centralized control fails due to an abundance of infrastructure or rolling stock disruptions or information system malfunctions. We test the performance of four rolling stock and two driver dispatching strategies in a microscopic simulation. Our test case is a part of the Dutch railway network, containing eleven stations linked by four train lines. We find that with the decentralized dispatching strategies, target frequencies of the lines are approximately met and train services are highly regular without large delays. Especially strategies that allow rolling stock to switch between lines result in a high performance