Data-Driven Empty Container Repositioning for Large Scale Railway Network With Fuzzy Demands

The Chinese rail network has a fast-growing number of containers in service to promote intermodal transport and attract more freight demand the railway for sustainable system. repositioning empty induces significant problem large-scale network, due spatial imbalance between supply demand. Traditionally, container (ECR) is solved with estimated parameters given distribution, which may difficult known decision maker. This article develops data-driven framework ECR based on large datasets available. employs machine-learning algorithms forecast supply/demand containers, draw identifying factors that can be hardly integrated into optimization models. Based those parameters, two models different modes are then proposed explore optimal plan. An weight coefficient introduced objective functions, minimizes total kilometers transported. by CPLEX after constraints conversion triangular fuzzy supply/demand. numerical results real-world case show solution method yield container-kilometer increase introduced, as some no longer follow shortest path, while final plan becomes practicable executable.