We live in a data-hungry world. Users of public transport systems now expect to be able to access live data about arrival times, transit connections, service disruptions, and many other types of status updates and reports at almost every stage of their journey. Studies suggest that providing real-time information on bus journeys and arrival times in this way encourages greater use of buses with beneficial effects for the bus service. In contrast, when use of buses decreases, transport experts suggest that this aggravates existing problems such as out-dated routes, bunching of vehicles, and insufficient provision of greenways or bus priority lanes. Each of these problems makes operating the bus service more difficult. Bus timetables become less dependable, new passengers are discouraged from using the bus service due to bad publicity, which leads inevitably to budget cuts that further accelerate the decline of the service.
Service regulators are no less data-hungry than passengers, requiring transport operators to report service-level statistics and key performance indicators which are used to assess the service delivered in practice against regulatory requirements on the quality of service expected. Many of these regulatory requirements relate to punctuality of buses, defined in terms of the percentage of buses which depart within the window of tolerance around the timetabled departure time; and reliability of buses, defined in terms of the number of miles planned and the number of miles operated.
With the aim of helping service providers to be able to work with models which can be used to analyse and predict on-time performance, the QUANTICOL project has connected a set of modelling and analysis tools into an analysis pathway, starting from system measurement data, going through data fitting, model generation, simulation and statistical model-checking to compute verification results which are of significance both to service providers and to regulatory authorities.
The QUANTICOL project is devoting more than the usual amount of effort to ensuring that our tools are user-friendly and easy-to-use. This is because we want our software tools to be used “in-house” by service providers because only then can service providers retain control over access to their own proprietary data about their service provision. With respect to ease-of-use in particular, making model parameterisation simpler is a crucial step in making models re-usable. Because vehicle occupancy fluctuates according to the seasons, with the consequence that buses spend more or less time at bus stops boarding passengers, it is essential to be able to re-parameterise and re-run models for different data sets from different months of the year.
It is also necessary to be able to re-run an analysis based on historical measurement data if timetables change, or the key definitions used in the evaluation of regulatory requirements change. Evidently, a high degree of automation in the process is essential, hence the project’s interest in an analysis pathway.
The QUANTICOL project has integrated quantitative analysis tools for data fitting, model generation, simulation, and statistical model-checking, creating an analysis pathway leading from system measurement data to verification results. The analysis pathway has been applied to service data obtained from Lothian Buses about the arrival and departure times of their buses on key bus routes through the city of Edinburgh.
More information: see the MultiVeStA site.
Authors: Andrea Vandin, Mirco Tribastone and Stephen Gilmore.