Improved Gravity Model for Network Passenger Flow Distribution Prediction in the Initial Stage of the Existing Urban Rail Transit Extension Lines

Abstract

［Objective］ The launch of extension sections for existing urban rail transit lines will affect passenger flow distribution throughout the network. It is necessary to propose a better model to meet the accurate prediction demand of network passenger flow during the initial stage of the extension section. ［Method］ An improved gravity model using the Follett method is proposed, and the opening of an existing line extension section in a city is selected as a case. The model parameters are calibrated in combination with the historical passenger flow characteristics to verify the feasibility and accuracy of the model. The actual case is used for prediction and compared with the prediction results of the improved double-constraint gravity model. ［Result & Conclusion］ The weighted average error of passenger flow distribution calculated by the improved gravity model with convergence of Follett method is 5.98 person-times, while most prediction results have an error of less than 10 person-times, which is even higher than the prediction accuracy of the double constrained gravity model.

Keywords

• urban rail transit
• passenger flow prediction
• gravity model