Probability-based bus headway regularity measure

Abstract

In frequently serviced bus routes passengers are more concerned about bus headway regularity than actual punctuality of bus arrival to the schedule. Buses arriving within very small (bus bunching) or very large headways are of particular concern and much less desirable. In this study, a time-point (stop) level probability-based headway regularity metric to measure bus service reliability is first formulated as a function of bus dwell time, number of stops into the trip, passenger activities (i.e. arrival, boarding and alighting) and expectation (or tolerance) of bus headways. The proposed metric is then applied to evaluate a Chicago Transit Authority bus route by using automatic vehicle location (AVL) data. It is found that headway regularity during a bus trip is closely impacted by dispatching headway. Furthermore, the time-point level service reliability declines as passenger activity levels increase or as the maximum passenger anticipated headway decreases (i.e. passengers become more demanding of frequent bus services). The case study demonstrates that the proposed probability-based headway regularity measure provides an operationable metric to transit agencies in terms of improving the transit service to meet passengers' expectation and thus increase ridership. Lastly, this study demonstrates another important application of the AVL/APC data.