Applying Virtual Modelling to Verify Control Systems Decision with Artificial Intelligence in Railway Transport

Abstract

This paper describes the applying of simulation modelling as a method for verifying the decision support system and forecasting railway traffic in the dispatching management of railway transport. This decision support system would be realised using artificial intelligence methods and on the training, stage requires an additional check of those decisions. For this purpose, a railway traffic model on a test railway section was implemented using a virtual model of the railway section's infrastructure. A virtual model of the railway section infrastructure is realised using the MPC-EL microprocessor-based centralization of arrows and traffic lights. The combination of a virtual railway section and a railway traffic model form a unified system. This system making a forecast train schedule based on initial data of solutions which are received from the decision support system. The verification of solutions obtained from the DSS would check this railway schedule practicability in real conditions. An additional feature of that system supporting the ability to work on workstations with the Elbrus processor, which in turn contributes substitution to the imported component base on the Russian Railways.