Transit network design considering link capacities

Abstract

Appropriate public transport systems are crucial in modern cities. Given the high costs that they represent and the impact they have on people's lives, effective tools are required to support their design. The Transit Network Design and Frequency Setting problem (TNDFSP) has been extensively studied in operations research. This problem consists of designing a set of public transport lines and a service frequency to each line. The main contribution of this work is to consider, for some links of the underlying network, a maximum ‘combined’ frequency among all the lines using that link and addressing the crowding issues that can result from that. These additional constraints intend to limit, for instance, the number of buses circulating in certain areas, in line with current urban design policies. A bi-objective memetic algorithm is proposed to solve the problem. The algorithm generates a set of non-dominated solutions that comply with the link-capacity constraints in 1 h of computing time. Additionally, alternative solutions are generated by designing the line plan without considering the link-capacity constraints and adapting the frequencies afterwards, to compare the two approaches. The algorithm is tested on an instance representing the bus network in the city of Utrecht, The Netherlands. The algorithm that takes into account the link-capacity constraints during the optimization process, generates better results. Moreover, the algorithm generates transit networks with less lines traversing the city centre, allowing higher individual frequencies for those lines. The algorithm could become an important tool for policy makers and transit operators, allowing the design of efficient transit systems that adjust better to contemporary urban requirements.