Virtual Coupling Train Marshaling Bi-level Optimization Model for Conflict Resolution at Junctions

Abstract:

In addressing the conflicts and throughput efficiency issues for virtual train formations at junctions， various marshaling coordination strategies were proposed. The train merging timing， sequence， formation composition， and speed sequence decisions were integrated into a bi-level optimization model. The upper-layer model optimizes the throughput efficiency， while the lower-layer model improves the virtual coupling formation quality. A bi-level nested particle swarm algorithm was designed， with the train merging time serving as an intermediate interactive variable between the two layers， enabling the convergence of train formations to different configurations based on environmental demands. Simulation results demonstrate that the inclusion of coordinated speed optimization can enhance the efficiency of the coupling process. The bi-level optimization model can increase the throughput efficiency of existing lines while ensuring the stability and balance of marshaling structures.