Designing a Model for Prediction the Suburban Daily Traffic Volume Using Adaptive Network Based Fuzzy Inference System

An ever increasing demand for using personal vehicles as means of transportation has turned the traffic congestion in to the most important problem in many metropolitan areas in the world. Environmental, social and economic impacts of traffic jams on human societies have urged the researchers to seek solutions to cope with this problem. Prediction of the daily volume of traffic is one of the solutions to the problem. The Traffic prediction helps the controllers to prevent the traffic congestion by taking appropriate decisions. In present study an Adaptive Network Based Fuzzy Inference System (ANFIS) has been used as one of the artificial intelligence-based methods in predicting the volume of suburban daily traffic in the routes connecting the provinces in Iran. To increase the reliability of the estimations, the data were randomly applied to the ANFIS in three modes of training, testing and checking: The training set was used in model design, while the testing and checking sets were used for validation of the model. The designed model predicts the traffic volumes with 90% precision. It can, therefore, be concluded that ANFIS is an appropriate tool for predicting the volumes of daily suburban traffic, and authorities are suggested using this model in their planning and decisions.