MODELING SCHEDULE DEVIATIONS OF BUSES USING AUTOMATIC VEHICLE-LOCATION DATA AND ARTIFICIAL NEURAL NETWORKS

Abstract

The establishment of the Advanced Public Transportation Systems program has encouraged bus transit operators to experiment with implementing automatic vehicle-location systems for real-time monitoring and supervision of operations. While the focus has primarily been on the implementation of technologies, such as automatic vehicle-location systems, it is necessary to experiment and develop advanced performance analysis and evaluation procedures that can assist in schedule planning and real-time service-control tasks. One potentially useful and effective approach to these tasks is system behavior modeling. In this study this method is used to model schedule behavior of buses on a route using schedule-deviation information. The primary objective of this study is to investigate the application of artificial neural networks, which have been shown to hold promise then applied to nonlinear dynamic system-modeling problems, for developing schedule behavior models. Models are developed using the schedule-deviation information obtained from Tidewater Regional Transit's automatic vehicle-location system. The time-series analysis approach is adopted for the development of schedule behavior models at the route level. The results of a case study are encouraging and demonstrate the usefulness of artificial neural network techniques, especially the Jordan networks and the Elman networks, for modeling schedule deviations of buses on a route.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 49 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.