Project: Measuring Demand for an Integrated Inter-Urban Public Transport Network

Reference: STP 14/6/16

Last update: 27/07/2004 10:58:06


The aim is to evaluate, with the aid of a cross-sectional demand model, "clean sheet" strategies for national and regional inter-urban networks, with particular reference to the proposition that a regular-interval, high-connectivity timetable could yield social, environmental and commercial benefits. There would be two measures of success: first, a model capable of routine application to estimating demand for travel by public transport between any pair of places or along any corridor, and second, a clear demonstration whether or not the policy-options under review should be adopted by Government and the public transport industry.


The model will estimate the demand for travel by public transport between any pair of places or along any corridor.


Passenger Transport Networks
49 Stonegate, York, yO1 8AW
01904 611187

Contract details

Cost to the Department: £52,849.00

Actual start date: 08 November 2001

Actual completion date: 17 December 2003


Measuring Demand for an Integrated Inter-Urban Public Transport Network
Author: Jonathan Tyler
Publication date: 08/12/2003
Source: Jonathan Tyler, Passenger Transport Networks, University of Leeds

The Philosophy and Practice of Taktfahrplan: a case-study of the East Coast Main Line
Author: Jonathan Tyler
Publication date: 01/11/2003
Source: Jonathan Tyler, Passenger Transport Networks, University of Leeds

Summary of results

    The key aim was to evaluate 'clean sheet' strategies for inter-urban public transport networks, with particular reference to the proposition that a regular-interval, high-connectivity timetable could yield social, environmental and commercial benefits. The two measures of success were, firstly, development of a model capable of routinely estimating demand between any pair of places, and secondly a clear demonstration whether or not this policy option should be adopted.
    The concern was that the institutional architecture of privatisation was not promoting strategic thought about the structure of services provided by the railway. The industry offered some financial support, a great deal of practical support and its endorsement of the proposed investigation. At the time the project was conceived a national timetable was not on the industry's agenda, even though it was a model commonplace in Europe. The consensus now is that the timetable-planning process introduced with privatisation appears to have failed, and the last few years have been a period of turbulence for many different reasons. The outcome has been renewed attention to timetabling, with
    Train Operating Companies becoming committed in principle to regular-interval services and the SRA playing a stronger role in achieving an integrated outcome. As a result, interest in our work is much greater than when we started, and the likelihood of its application is therefore greatly enhanced.

    Four principal themes run through the research. These are:-
    1) identification and assessment of 'ideal' inter-urban public transport networks;
    2) enhanced understanding of the determinants of the catchments of stations;
    3) development and calibration of a fresh cross-sectional rail demand model; and
    4) specification and evaluation of a national, regular-interval, coordinated timetable.

    Two approaches to identifying an 'ideal' network were developed. The first employed simulated annealing, a procedure for optimally locating nodes, together with a program to convert the nodes into a credible network. It was found however that the results were highly sensitive to the input assumption about decay-elasticity. We also employed an aggregation procedure based on size and proximity of population units and formulated a nationwide urban hierarchy. This work produced provisional material that has served well in our judgements about service patterns in the timetabling exercises. An important conclusion is that many suburban areas are poorly connected with the national system. Because station catchments have been relatively little researched we also analysed postcoded records from customer surveys. The results raise the possibility that people with an above- norm propensity to travel by train locate nearer stations, which has implications for network-planning.

    We adopted a new model developed by an EPSRC4unded. doctoral student at ITS. The strength of this cross-sectional model is that it incorporates the access effects, rather than handling them indirectly, using a Geographic Information System to generate spatial population and road network data. It is calibrated against observed journeys for pairs of stations and estimates plausible parameters, including the proportion of travellers using each of several competing stations. The model is very suitable for testing changes that alter station choices and catchments, unlike that now used for demand forecasting which cannot handle a substantial recasting of a timetable.

    In order also to incorporate explicitly the impact of regular-interval timetables both stated-preference and revealed-preference approaches were adopted. From passengers' choices between hypothetical timetables with different characteristics elasticities were derived for the degree to which trains run at the same minutes past each hour, for the evenness of intervals and for timings at 'memorable' minutes. These findings were supported by revealed values in the cross-sectional model. Through conversations with Swiss organisations we gained a thorough understanding of the principles and methodology of a Taktfahrplai~ in order to construct draft timetables with the Viriato software. Tight coordination based on coherent rules about the relationships between services achieves a high
    degree of connectivity, and presentation is made simple by operation in unvarying hourly cycles. The timetables were rooted in real operational data, with the help of the industry. Case-studies have been worked through for the Birmingham area, South West England, the East Coast Main Line, the Barnstaple Branch and Scotland. These exercises have demonstrated that there is no intrinsic reason why public transport in Britain cannot be operated on the same principles as the Swiss system. The model was used to forecast the change in patronage following recasting of the East Coast time table. The introduction of a Taktfahrplan would increase rail traffic. The aggregate appraisal found
    that revenue and user benefits would amount to about £23 million. Given the low market share that rail holds on non-London routes the large benefits they would yield could have implications for the relative priority of projects in the context of the Government's environmental and social objectives.

    Collaboration between all the parties has worked extremely well. The staff of ITS and PTN, the Project Manager, have had excellent relationships with Network Rail and ATOC, particularly through being given extensive access to data. Eden Business Analysis provided knowledgeable advice on many technical matters. SMA contributed the Viriato software without which the task would have been impossible and has been a rich source of wisdom and encouragement. The Strategic Rail Authority supported the application and joined the Management Group as an observer. Our work has benefited substantially from briefings by senior staff of Train Operating Companies.
    Meetings of the Management Group were held at appropriate intervals. Because this is a topic of public interest the research has been accompanied by extensive activity to learn from a wider audience.

Departmental Assessment Status: Assessed by FIT Programme Advisory Group 15 July 2004. Allocated score 7/10 - Very good project, achieved targets and objectives, good quality research work, outputs with commercial potential.