Augmented Network Based Urban Transit Assignment Model and Solution Algorithm under Congestion
Ming Zhong,Bingfeng Si,Huijun Sun,Ziyou Gao
2010-01-01
Abstract:Passenger flow assignment for urban transit network is relatively complicated due to the complexity of its super-network structure including both physical routes, stations/lines, and many other factors, such as waiting time, travel time, congestion and transfers, influencing passengers' route and line choices. In the past decades, various transit assignment models have been proposed to simulate passengers' behaviors. However, the existing approaches generally treat passengers' combined route/line choices as the problem for which complicated formulations and greedy algorithms are required to determine an optimal subset of attractive common lines at stations and then use them as one of aggregate choices. In general, it is very difficult to apply these approaches to a large-scale, realistic transit network. In this paper, a new technique called augmented network is proposed to represent urban transit system. The proposed approach is capable of eliminating the complexity of common-line problems required in traditional transit passenger flow assignment. Through this augmentation technique, urban transit system can be simplified and represented by an augmented network - it then behaves like a simple network and can be used directly as a generalized network for traffic assignment purpose. The congestion at station and in vehicle can also be easily incorporated. On the basis of these, an equilibrium assignment model with asymmetrical travel cost is presented and the corresponding solution based on a conditional shortest path method is also proposed. Finally, a numerical example is provided to illustrate our approach.