EVALUATION OF ARTIFICIAL NEURAL NETWORK APPLICATIONS IN TRANSPORTATION ENGINEERING

The increased interest in artificial neural networks (ANNs) seen in government and private research as well as business and industry has included relatively little activity in transportation engineering. The position that ANNs, as a branch of artificial intelligence, hold in the transportation engineering field is discussed, including the differences between ANNs and biological neural networks and expert systems, respectively. The characteristics of ANNs in different fields are discussed and summarized, and their potential applications in transportation engineering are explored. A case study of trip generation forecasting using one traditional method and two ANN models is presented to show the application potential of ANNs in transportation engineering. The results of each method are compared and analyzed, and it is concluded that the potential for using ANNs to enhance both software and hardware in transportation engineering applications is high, even in comparison with expert systems and other types of artificial intelligence technique.