Using AVL Data to Improve Transit On-Time Performance

This paper describes an approach for improving on-time performance at transit agencies. It takes advantage of the schedule adherence information from an AVL system. A methodology that can be used to update the bus timetables by using AVL schedule adherence data is described. Using statistical analysis, the main goal is to maximize the density area of the on-time performance range. From this distribution, the optimal value is obtained and used to update the times in the timetables. Then, a comparison process is used to assess the on-time performance improvements. In addition, a simulation process is presented to provide a different perspective than the statistical methodology. This approach also presents possibilities for further ontime performance improvements. To demonstrate the applicability of this research, a case study using data from Miami-Dade Transit is included. The on-time performance calculations for Routes 99 and 57 also are presented. Introduction To passengers, schedule adherence is a matter of service quality. From the service provider perspective, schedule adherence reflects the quality of the service plan (the schedule) and the operations control (Furth et al. 2003). Researchers have long noticed the importance of schedule adherence information contained in Automatic Vehicle Location (AVL) systems. Lee et al. (2001) studied the effect of an AVL system on schedule adherence and operator behavior and willingness to Journal of Public Transportation, Vol. 14, No. 3, 2011 22 keep on schedule. In addition, Hammerle et al. (2005) pointed out that some transit agencies would like to use Automatic Passenger Counter (APC) and AVL data to inform service planning and management and ultimately provide more reliable service. Methods for extracting information from these data were developed to compute service reliability indicators. Also, some schedule adherence properties were observed and reported in their research. These studies show a general interest in improving schedule adherence. It is important to clarify the difference between schedule adherence and on-time performance. Schedule adherence refers to the difference between real time and scheduled times of arrival or departures times, usually presented in minutes. Ontime performance, on the other hand, is a percentage value used to indicate buses arriving or departing late, on time, or early. Depending on the AVL system and the transit agency, on-time performance can be calculated using arrivals, departures, or possibly a combination of both. AVL systems are computer-based vehicle tracking systems that function by measuring the real-time position of each vehicle and relaying this information back to a central location. Many researchers also see the potential uses of analyzing AVL or APC data to improve service quality. A study that uses data from Tri-Met in Portland, Oregon, shows that scheduling can be improved through performance monitoring using AVL data and that very useful information has been retrieved (Kimpel et al. 2004). Shalaby and Farhan (2004) made efforts to use AVL and APC data to develop a bus travel time model capable of providing real-time information on bus arrival and departure times to passengers (via traveler information services) and to transit controllers for the application of proactive control strategies. One continual question asked by researchers is how to use AVL data to improve on-time performance. The importance of on-time performance to both the transit customer and the transit providers has been discussed in many research projects. For instance, New York City transit established a customer-oriented bus performance indicators program to measure on-time performance. The program contains two schedule adherence indicators that measure different aspects of service performance: route on-time performance and service regularity. The purpose of this program is to measure the quality of service experienced by the customer (Nakanishi 1997). This research attempts to fill the gap in the understanding of AVL data and presents another perspective on how the data can be used to improve on-time performance. With the availability of AVL data, it is possible to improve on-time