Accommodating user preferences in the optimisation of public transport travel

• Main Author: Wu, Qiujin
• Published: Nottingham Trent University 2006
• Subjects: 388.40941
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.441487

Currently, most of the big cities are facing the traffic congestion problem. Encouraging more people to use public transportation is considered as one of the best solutions (Hartley and Bargiela, 2001). To provide travellers with individual travel advice for journeys, today's Advanced Traffic Information Systems require efficient algorithms that are able to find the `best' route(s) for the public transport users according to their preferences. The aim of this research is to develop optimal and efficient time-dependent shortest path(s) algorithms which can accommodate public transportation users' preferences such as `minimum travel time', `minimum number of bus-changes', and `minimum walking time'. Two solutions are provided and compared in this report: the objective-oriented shortest path algorithms and the K-shortest paths (KSP) algorithm. Three different objective-oriented shortest path algorithms are used to accommodate three different user preferences separately. To find a route with `minimum travel time' (earliest arrival time or latest departure time), a forward/backward search algorithm is developed. The combination of the minimum transfer path search algorithm and the forward search algorithm is sufficient to find the path with the `minimum number of buschanges' (Huang and Peng, 2002). With a slight modification to the forwardibackward search algorithm, the algorithm can find a route that has `minimum walking time'. The K-shortest paths algorithm considers the three different user preferences simultaneously. It computes a reasonable number of ranked shortest paths, with the ultimate `most optimal' path being selected by consideration of the preferences. The experimental results based on the public transportation network of Nottingham City show that the objective-oriented shortest path algorithms work efficiently but can only optimise one single user preference. Using the KSP algorithm to accommodate user preferences is feasible. However, the execution time of the KSP algorithm is still long. A heuristic method is thus proposed to reduce the execution time by solving the `overlap ratio' between the k shortest routes. Experiment results are presented in Chapter Six