This paper presents the impacts of the implementation of automatic vehicle location (AVL) and computer aided dispatching (CAD) deployed in 1998 by the Tri-Met transit agency in the Portland, Oregon metropolitan area. The study established the impacts of the system through analysis of data collected over 10 weekdays prior to the implementation and a 10-day period shortly after the system began operation. The 8 routes surveyed in the study included a cross section of the types of routes in operation by Tri-Met. For comparison purposes, trips from data collected before system implementation were matched to the corresponding trip in the data collected by the new system.

Comparing the before and after data revealed several benefits of the AVL and CAD systems. These benefits include a 9.4 percent improvement in on-time performance measured at the final destination of the routes under study; improvements at earlier points on the route were likely higher. The variability in the headways between buses decreased by 5 percent after the implementation of the improvements. No significant change was measured in the average run times for buses along the routes, with run times remaining about 1 percent longer than their scheduled values. The average coefficient of variability for bus run times did improve by 18 percent however, and no route experienced an increase in run time variability. The benefits indicated by the comparison of before and after data are consistent with the improved control available to transit supervisors after the implementation of the AVL and CAD systems.

Automatic passenger counters were available on some of the buses, and this data was matched to the corresponding before and after data sets. This provided information on passenger loads for 820 of the approximately 3,000 bus trips surveyed during the study. A quantitative model developed using the 820 data sets helped assess the influence of various other factors on the change in vehicle run times between the before and after periods. The model incorporated departure delays, route length, the number of stops, boardings, and alightings, scheduled headways, peak periods and the presence of the AVL/CAD system. Results of the modeling effort indicate that the influence of the AVL/CAD system provides a 3.14 percent improvement in run time. This improvement in run time allows the transit agency to save operating costs.The average run-time measured in the study increased slightly between the before and after periods (from 45.4 to 46.0 minutes). The model results indicate that the improvement in vehicle run times due to the AVL/CAD system was offset by other influences such as the measured increases in the number of stops made, scheduled headways, and departure delays which increased the run times of vehicles. This means that the AVL/CAD system allows the transit agency to provide the same level of service to a greater number of travelers with the same equipment, increasing the effective capacity of the bus system.

The authors note that the benefits identified in this study were measured shortly after the AVL and CAD systems began operations. Additional benefits are likely to accrue as managers, dispatchers, field supervisors, and service planners become more accustomed to the system.