A Study on Large Scale Railway Origin-Destination Traffic Volume Distribution

This paper mainly discusses the problems in large-scale traffic volume distribution, which includes the selection of distribution models, the data preconditioning and the valuation of large sparse matrixes for certain kinds of goods, etc. Analysis and computation show thatthe improved increasing coefficientmodel and physical analogue model are suitable to the prediction of large-scale railway origin-destination traffic volume distribution. By using numerical decomposition and superposition of the predicted transportation flux, the planned data can be processed with higher precision. Although fromthe point of viewof computation it is possible to assign tiny values to the zero elements of the origin-destination matrixes, it conceals the actual unbalance in local area networks.