Several issues are discussed to realize advanced train rescheduling systems in the next generation. We clarify that we should take an approach to construct them as an aggregation of several subsystems and insist that the subsystems be divided not by their functions but by considering their objectives. We also insist that some of the issues which need further studies are; the methods to evaluate train rescheduling based on more direct criteria and those by considering the robustness of the rescheduled plans against risks that would occur.

Copyright (c) 2004 Railway Technical Research Institute