Forecasting the Travel Demand for Tehran-Mashhad Rail Route

Travelling is an important aspect of human life. Although various technologies invented to reduce the travel demand, there is an increasing demand for urban and suburban travels. Obtaining transportation facilities to satisfy this demand is one of the most important roles of transportation decision makers. To establish an appropriate policy to adjust the travel supply and demand, the future demand should be forecasted. Railway is one of the safest, cleanest and cheapest modes of travel. Also emerging technologies make it one of the fastest travel modes. Therefore the rail system has a significant travel share among other modes and forecast of its demand is very important. This paper presents some models to forecast the peak (March & September) and off-peak rail travel demand for Tehran-Mashhad which has the most rail traffic in the Iran railway system. The peak travel demand for Tehran-Mashhad railway usually exceeds the rail capacity and the level of insatiable demand is unknown. Constructing the travel demand forecasting model requires historical data of dependent and independent variables. Prior researches have used the number of passengers travelled between Tehran and Mashhad by the rail system instead of real peak demand because of unavailability of data for travel peak demand as dependent variable of model. Identifying the factors affecting travel peak and off-peak demand, this research presents Multiple Regression (MR), Artificial Neural Network (ANN) and two Time Series models produced to forecast the travel demand for Tehran-Mashhad railway. MR and ANN are able to forecast the rail travel demand with more than 95% accuracy. The models estimation of unsatisfied demand for rail travel in peak periods is about 5.5%.