Track: Transport and Traffic
Forecasting for passenger demand in a new railway service is a challenging activity. There are many variables that could affect the number of passengers that will use railway services and many of those variables may be unique to each specific situation. In this study, we present a study that attempts to forecast railway passengers for a newly constructed railway line of Makassar – Parepare in South Sulawesi, Indonesia. We propose a combination of the System Dynamics and Artificial Neural Networks (ANN). The system dynamics is a simulation method that is used to generate a set of training data for the ANN. The initial data were collected through a survey of willingness to use and willingness to pay among the selected potential passengers. Different price scenarios are also investigated in the model. Our study shows that with an optimistic scenario, there are potentially around 8000 – 9000 train tickets sold weekly. However, as the study is conducted prior to the railway operations, the results need to be reviewed and adjusted once it is operated and sufficient initial data is available.
Passenger Demand Forecasting, Railway Passengers, System Dynamics, Artificial Neural Network, Indonesia.