Towards Developing a Travel Time Forecasting Model for Location-Based Services: A Review

Abstract

Travel time forecasting models have been studied intensively as a subject of Intelligent Transportation Systems (ITS), particularly in the topics of advanced traffic management systems (ATMS), advanced traveler information systems (ATIS), and commercial vehicle operations (CVO). While the concept of travel time forecasting is relatively simple, it involves a notably complicated task of implementing even a simple model. Thus, existing forecasting models are diverse in their original formulations, including mathematical optimizations, computer simulations, statistics, and artificial intelligence. A comprehensive literature review, therefore, would assist in formulating a more reliable travel time forecasting model. On the other hand, geographic information systems (GIS) technologies primarily provide the capability of spatial and network database management, as well as technology management. Thus, GIS could support travel time forecasting in various ways by providing useful functions to both the managers in transportation management and information centers (TMICs) and the external users. Thus, in developing a travel time forecasting model, GIS could play important roles in the management of real-time and historical traffic data, the integration of multiple subsystems, and the assistance of information management. The purpose of this paper is to review various models and technologies that have been used for developing a travel time forecasting model with geographic information systems (GIS) technologies. Reviewed forecasting models in this paper include historical profile approaches, time series models, nonparametric regression models, traffic simulations, dynamic traffic assignment models, and neural networks. The potential roles and functions of GIS in travel time forecasting are also discussed.