Department of Industrial Engineering

Permanent URI for this community

https://scholar.sun.ac.za/handle/10019.1/104

Browse

Browsing Department of Industrial Engineering by browse.metadata.advisor "Andersen, Simen Johann"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Bus route design and frequency setting for public transit systems
    (Stellenbosch : Stellenbosch University, 2022-04) Husselmann, Gunther; Van Vuuren, Jan Harm; Andersen, Simen Johann; Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering.
    ENGLISH ABSTRACT: The availability of effective public transport systems is increasingly becoming an urgent problem in urban areas worldwide due to the traffic congestion caused by private vehicles. The careful design of such a transport system is important because, if well designed, such a system can increase the comfort of commuters and ensure that they arrive at their destinations timeously. A well-designed public transport system can also result in considerable cost savings for the operator of the system. The problem considered in this dissertation is that of designing three mathematical models for aiding a bus company in deciding upon efficient bus transit routes (facilitated by the first two models) and setting appropriate frequencies for buses along these routes (facilitated by the third model). The design criteria embedded in the first model (for designing bus routes) are the simultaneous pursuit of minimising the expected average passenger travel time and minimising the system operator’s cost (measuring the latter as the sum total of all route lengths in the system). The first model takes as input an origin-destination demand matrix for a specified set of bus stops, along with the corresponding road network structure, and returns as output a set of bus route solutions. The decision maker can then select one of these route sets subjectively, based on the desired trade-off achieved between the aforementioned transit system design criteria. This bi-objective minimisation problem is solved approximately in three distinct stages — a solution initialisation stage, an intermediate analysis stage and an iterative metaheuristic search stage during which high-quality trade-off solutions are sought. A novel procedure is introduced for the solution initialisation stage aimed at effectively generating high-quality initial feasible solutions. Two metaheuristics are adopted for the solution implementation, namely a dominance-based multi-objective simulated annealing algorithm and an improved non-dominated sorting genetic algorithm. The second model is a novel approach towards establishing high-quality bus routes resembling a reference set of bus routes (typically the currently operational bus routes) to varying degrees, providing the decision maker with bus route design alternatives that may be implemented incre mentally in order to limit the disruption experienced by passengers in the bus transit network. The objectives pursued in this model are the simultaneous minimisation of the expected aver age passenger travel time and the minimisation of a reference-route-to-design-route similarity measure. The second model takes the same input as the first model above, with the addition of a reference route set with which to compare alternative design routes in terms of similarity, and provides as output a set of trade-off solutions according to this model’s design criteria. The same three-stage approximate solution methodology described above is adopted for this model, and the same two metaheuristic implementations are utilised to solve instances of this new model. In the third model, high-quality bus frequencies are sought for each bus route in pursuit of min imising the expected average travel time for passengers (including waiting time, transfer time and travel time) and simultaneously minimising the total number of buses required by an operator to maintain the specified frequencies. The third model takes as input all the data required by the first model, along with a route set for which frequencies should be set, and returns as output a set of bus frequencies at which buses should operate along the various routes, based on a de sired trade-off between the aforementioned two design criteria. The solution approach adopted for this bi-objective minimisation problem again conforms to the three aforementioned distinct stages, with the exception that only a non-dominated sorting genetic algorithm is designed for solving it. The first and third models are finally applied to a special case study involving real data in order to showcase the practical applicability of the modelling approach.