Integrating Passengers' Assignment in Cost-Optimal Line Planning

Finding a line plan with corresponding frequencies is an mportant stage of planning a public transport system. A line plan should permit all passengers to travel with an appropriate quality at appropriate costs for the public transport operator. Traditional line planning procedures proceed sequentially: In a first step a traffic assignment allocates passengers to routes in the network, often by means of a shortest path assignment. The resulting traffic loads are used in a second step to determine a cost-optimal line concept. It is well known that travel time of the resulting line concept depends on the traffic assignment. In this paper we investigate the impact of the assignment on the operating costs of the line concept. We show that the traffic assignment has significant influence on the costs even if all passengers are routed on shortest paths. We formulate an integrated model and analyze the error we can make by using the traditional approach and solve it sequentially. We give bounds on the error in special cases. We furthermore investigate and enhance three heuristics for finding an initial passengers’ assignment and compare the resulting line concepts in terms of operating costs and passengers’ travel time. It turns out that the costs of a line concept can be reduced significantly if passengers are not necessarily routed on shortest paths and that it is beneficial for the travel time and the costs to include knowledge on the line pool already in the assignment step.