A New Railway Line Planning Model Considering Multinomial Logit-based Traffic Assignment

초록

Most line planning models in railway network assume that the demands for each train type are in advance given from a traffic assignment model such as `system split'. System split is a procedure to distribute passenger demands over an entire network. This distribution is based on a value assigned to a route which is primarily determined from the total travel time, the number of transfers and maximum headway (or minimum frequency) of a route. However such stepwise approaches have a critical shortcoming that the line frequencies assumed in traffic assignment does not coincide with the line frequencies found by line planning model. Such inconsistency results in inaccurate demand and line frequencies. We present a uni？ed line planning model considering stochastic traffic assignment in order to overcome the inconsistency between line planning and traffic assignment.