Abstract:
The most common problems for passengers in any bus transportation network are those of delays, and uneven service and inconvenience caused by bus clustering. The cause of these problems is often an inadequate timetable, and therefore finding ways to improve the timetabling process is of high importance. The focus of this thesis is on applying optimisation and simulation techniques to the timetabling process, in order to reduce the amount of bus clustering which occurs. First, the causes of delays and bus clustering are explored, as well as potential solutions to these in both the form of timetable improvements and behavioural strategies for buses in the network. A key factor identified is the spacing of the arrivals at stops due to the timetable. The advantages of periodic timetables are also discussed, and these are decided to be the preferred choice. An optimisation model to target bus clustering is formulated. The objective function used aims to maximise the spacing between arrivals in the generated feasible timetable, while the constraints ensure a minimum separation. A simulation model is then created in order to test the generated timetables. Of particular importance are the introduction of random travel and waiting times and the ability to modify bus behaviour. The optimisation and simulation models are integrated using a user interface in order to create a tool which can more easily be used to test multiple bus networks and configurations of these networks. Finally, this integrated package is used to test two network case studies, and the performance of the timetables for these networks is investigated for different network parameters.
