Optimization on section- and period -based differentiated tolling strategies using route selection modeling

Objective Excessive discounts on specific expressway sections may reduce total revenue with section- and period -based differentiated tolling. This study proposes a route selection model that integrates multi-dimensional cost analysis. The model quantifies the relation between vehicle route selection and toll revenue, improving existing tolling strategies. Method An improved Logit discrete selection model was developed based on truck drivers' route selection behavior. Economic cost, time cost, comfort cost, and safety cost were included in the utility function. Route selection probabilities were then estimated. An adaptive step stochastic gradient boosting algorithm was proposed through dynamically adjusting the step size to improve solution accuracy and convergence speed. Data from nine expressways, which implemented section- and period -based differentiated tolling, in Fujian Province were used for model application. The dual optimization scheme was designed, including discount rate adjustment and preferential section replacement. The discount levels remained unchanged for price-sensitive trucks of category Ⅵ. Daytime and nighttime discounts for trucks of category Ⅳ and category Ⅴ were reduced. The alternative preferential sections were proposed for loss-making sections. The effect of attracting traffic volume was evaluated using the improved Logit model. Result The recommended scheme attracts the average of 727 vehicles per day from competing truck routes. It also reduces inefficient discounts. The annual road operator revenue increases by 16.34 million CNY, i.e., 14.50 million CNY comes from toll rate adjustment, and 1.84 million CNY comes from section replacement. The total network revenue increases by 5.6% compared with that using policies before adjustment. Conclusion The improved Logit discrete selection model provides scientific basis and practical guidance for the optimization of the proposed differentiated tolling strategy.