Optimizing The Paths Of Trains Formed At The Loading Area In A Multi-loop Rail Network

With the continuous construction of rail lines in recent years,China's railway network features a multi-loop structure,and higher density and accessibility.In this case,the number of potential paths for each cargo flow has increased rapidly,which results in the problem of path selection for cargo flows.This paper focused on the transportation process of direct train from loading area to unloading area in the multi-loop rail network and studied the problem of optimizing the train path,which is of great practical significance in rationally allocating the rail network resources,improving the transportation efficiency and increasing revenue for railway company.The main research of the dissertation included the following contents:(1)The basic principle of train routing optimization and the theories of organizing direct train formed at loading area are analyzed.The formation mechanism of multi-loop rail network is introduced.And then three cases of the direct cargo flows' route selection are discussed according to the capacity of the multi-loop rail network.(2)In view of the insufficient capacity for a single arc in each loop to deliver all the cargo flows through the shortest path,a mathematical model(Model I)bounded with arc capacity constraint is established.(3)Considering that there may exist a capacity bottleneck in the network and can not meet all the cargo flows,the other path selection model(Model II)for direct cargo flows is established,which has taken the capacity bottleneck into account.(4)To solve the routing problem of direct train formed at loading area in a multi-loop rail network.,three optimization methods based on the former two models are designed,namely Model I-branch and bound,Model II-branch and bound and Model II-genetic algorithm,respectively.Computational experiments are carried out then.Same results can be obtained by Model I-branch and bound and Model II-branch and bound.The error of the result obtained by Model II-genetic algorithm is about 0.9% when the rail network capacity is sufficient,and the error obtained by Model II-genetic algorithm is about 0.7% when rail network capacity is insufficient.The performance of the three methods in the four groups of experiments indicates that Model I-branch and bound is recommended when the rail network capacity is sufficient,Model II-branch and bound is recommended when the rail network capacity is insufficient while the scale of the problem is small,and Model II-genetic algorithm is recommended to solve large scale problem when the rail network capacity is insufficient.