This paper addresses a railway station real-time dispatching problem that integrates the rescheduling, routing, and shunting issues related to the movement of trains, passenger coaches, locomotives and shunting locomotives at a busy station. We propose a novel microscopic agent-based simulation for solving this problem efficiently. A transformed network that reflects the particular station layout is derived specifically. The four pillars (agents, agents' environment, agents' relationship and methods of interaction) that form the agent-based model are defined exclusively for our simulation model. A clock-based activation which is implemented to assign the distinctive operation to each agent and a cell-based occupation monitor system is proposed to schedule the conflict-free running at a microscopic level of movable units. Three typical rules are implemented and tested to improve the efficiency of the simulation process. Computational experiments are performed on a realistic railway instance to validate the effectiveness of the simulation model with the three rules. The preliminary results show that our simulation system has a very fast execution time (about 15 sec).
He, B., Chen, P., D'Ariano, A., Chen, L., Zhang, H., Lu, G. (2020). A microscopic agent-based simulation for real-time dispatching problem of a busy railway passenger station. In 2020 IEEE 23rd International Conference on Intelligent Transportation Systems, ITSC 2020 (pp.1-7). Institute of Electrical and Electronics Engineers Inc. [10.1109/ITSC45102.2020.9294651].
A microscopic agent-based simulation for real-time dispatching problem of a busy railway passenger station
D'Ariano A.;
2020-01-01
Abstract
This paper addresses a railway station real-time dispatching problem that integrates the rescheduling, routing, and shunting issues related to the movement of trains, passenger coaches, locomotives and shunting locomotives at a busy station. We propose a novel microscopic agent-based simulation for solving this problem efficiently. A transformed network that reflects the particular station layout is derived specifically. The four pillars (agents, agents' environment, agents' relationship and methods of interaction) that form the agent-based model are defined exclusively for our simulation model. A clock-based activation which is implemented to assign the distinctive operation to each agent and a cell-based occupation monitor system is proposed to schedule the conflict-free running at a microscopic level of movable units. Three typical rules are implemented and tested to improve the efficiency of the simulation process. Computational experiments are performed on a realistic railway instance to validate the effectiveness of the simulation model with the three rules. The preliminary results show that our simulation system has a very fast execution time (about 15 sec).I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
