2017 年 2017 巻 p. 168-173
Japanese railway network especially in metropolitan area has become more complex or overcrowded. Railway companies must recovery stable operations when any accidents or troubles happen in their railway lines. However, computational supports for resolving train rescheduling problems are limited due to complexity of these problems. To make computational supports expanded, it is important to predict the behavior of trains exactly. In this research, we applied a theoretical traffic flow model to a real line in Tokyo. We observe that area of good parameters shifts according to time period or simulated train. Finally we conclude that the relation between given parameters and the accuracy of predictions is so complex that we cannot improve that accuracy only by enlarging the number of cells that are used in the employed Cellular Automata model.