A geo-clustering approach for distance-based toll pricing optimization using progressive graph convolutional networks to minimize congestion

Abstract

Toll gates play a crucial role in funding road infrastructure improvements and maintaining safe and efficient roads. However, traditional toll collection methods often lead to traffic congestion and increased travel times due to vehicle stoppages. This study introduces a novel Geo-Clustering Approach for Distance-Based Toll Pricing Optimization Using Progressive Graph Convolutional Networks (GC-DTP-MC-PGCN) to minimize congestion and improve toll pricing efficiency. Conventional static pricing models use fixed rates. In contrast, this approach dynamically adjusts toll rates based on predicted route choices and travel distances. Initially the input data is collected from the Austin Mopac Express Lane Prices dataset. A Fast Continual Multi-View Clustering (FCMVC) algorithm is employed to segment the tolling zones based on location coordinates for distance-based toll pricing. Progressive Graph Convolutional Networks (PGCN) is then utilized to predict user route choices and estimate the pricing based on the distance traveled within the identified tolling zones. The proposed technique is executed in Python and evaluated through performance metrics like Mean Squared Error (MSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE). The proposed method achieves superior performance with a MSE of 0.05 and RMSE of 10 at the 1-h prediction horizon, outperforming existing methods for example, using a Multi-Layer-Neural network (DCP-MUN-MLNN) to dynamically optimize congestion pricing in multi-region urban networks; using deep learning and fuzzy evaluation for toll-gate congestion prediction in China's freeway network (CPT-DF-GCN); and using deep learning, queuing theory, and differential evolution (CHT-DQD-RNN) to control highway toll stations. The proposed method achieves a calculation time of only 50 min for 1-h predictions, enabling rapid toll price adjustments and enhancing system responsiveness. These advancements contribute to more efficient traffic management and improved road network sustainability.