Metro Transit Service Reliability Measures Assessment

The objectives of this project are to expand the definitions of transit service reliability, determine which ones are applicable to Metro Transit's services, and provide a proof of concept of how these measures could be applied using available operations data. We first reviewed common service reliability measures that are used for bus and light rail services both in the United States and internationally. We identified 18 common service reliability measures and categorized them into four groups (see Appendix A). Four service reliability measures, On-Time Performance (OTP), Excess Wait Time (EWT), Gap Assessment (GA), and the 90th percentile of lateness (LA90), were finalized for further evaluation. We applied the finalized service reliability measure on 13 high frequency routes (route 2, 4, 5, 6, 10, 17, 18, 19, 21, 54, 64, 84, and 515) and 2 LRT lines (901-blue and 902 - green lines) using transit operations data from Jan. 2014 to Apr. 2016. The following selection criteria are applied to evaluate how the 4 selected measures are applicable to transit services and how they might be used most effectively by Metro Transit. 1. Is the service reliability measure easy to interpret? 2. Is data available for both historical reporting and real-time applications? 3. Can the metric provide a different perspective as compared to current OTP? 4. Can the metric support best operational practices? OTP is the current key performance index used by the Metro Transit. It will continue to be used as a baseline measure for service reliability. In addition to OTP, GA, and LA90 were selected for implementation. GA is a customer focused headway adherence measure. It could be easily implemented based on existing transit operations data. In addition to the OTP, the LA90 measure describes the typical late range of a transit service. It indicates 'how badly' the transit services are not meeting the OTP threshold. The excess wait time (EWT) is not recommended for implementation because it is difficult to interpret and the computed results are unable to provide a clear indication of operations performance. This study demonstrates how the identified measures could be applied using available operations data currently collected by Metro Transit to improve quality of service.