01519338

Component

Schedule adherence (SA) is one of the primary quality-of-service metrics reported by transit agencies in the United States. SA is a set of summary statistics regarding the difference between actual and scheduled arrival times at designated time points. Measures are calculated from archived CAD/AVL data and schedules; statistics include percentage of on-time arrivals, average and standard deviation. The time difference from actual and scheduled arrival at a time point is schedule deviation (a negative deviation is late, positive is early). It is understood that dwell time, traffic conditions, counts of boarding and alighting passengers, bus load, construction and weather are among the major factors influencing schedule adherence. The effect of driver performance is not generally considered. In this study, a statistical model isolates the effect of drivers on staying on schedule to the next time point. From a passenger perspective, predictability of arrival is essential to a satisfactory transit experience. Thus, some measure of is a critical component of service quality. Some analytics platforms present graphs of deviation distributions by time point on a route. The authors present a method for deeper analysis using more advanced statistical models -- the authors simultaneously regress covariates on means and dispersions of deviation distributions, thus accounting for driver’s contribution to the average deviation from schedule and the variability of deviation. A skilled driver arrives on time on average (is accurate), but also minimizes variation (is precise). The authors demonstrate this method with CAD/AVL data from a mid-sized US transit agency.

This paper was sponsored by TRB committee AP010 Transit Management and Performance.

01503729

14-3910

English

12p

2014

Transportation Research Board 93rd Annual Meeting

Digital/other

Figures; References; Tables

Automatic vehicle location; Boarding and alighting; Driver performance; Dwell time; On time performance; Quality of service; Schedules and scheduling; Statistical analysis

United States

Planning and Forecasting; Public Transportation; I72: Traffic and Transport Planning

Transportation Research Board Annual Meeting 2014 Paper #14-3910

TRIS, TRB, ATRI

Jan 27 2014 3:21PM