An Agent Based Approach for Train Traffic Control

Abstract

The Traffic Control department of ProRail manages the train traffic on the railway network, especially in cases of calamities. Train Traffic Controllers (referred to as TRDLs) are the people at this department that are responsible for safely guiding everything in an assigned area. A TRDL achieves this mainly by controlling signs, switches and railway settings. These TRDLs are supported by an Automatic Railway Setting (referred to as ARI). The main shortcoming of the ARI is that many real world traffic management problems are not supported by the functionality. The goal of this research project is to specify, design and implement a multi-agent system within ProRails distributed simulated environment to support TRDLs in their work that is not handled by the ARI. The specification and design of this multi-agent system were modeled using the Prometheus methodology. The implementation of the MAS is created with the OO2APL Java library. I was able to create a working proof of concept of a multi-agent system within a distributed simulation environment that notifies when a Train Handling Documents (referred to as, TAD(s)) should be applied to ProRail. The system not only does it at the decision point, but it gives the TRDLs a heads up to possibly prepare a TAD.