Abstract:This article adopts a family of surrogate‐based optimization approaches to approximate the response surface for the transportation simulation input–output mapping and search for the optimal toll charges in a transportation network. The computational effort can thus be significantly reduced for the expensive‐to‐evaluate optimization problem. Meanwhile, the random noise that always occurs through simulations can be addressed by this family of approaches. Both one‐stage and two‐stage surrogate models are tested and compared. A suboptimal exploration strategy and a global exploration strategy are incorporated and validated. A simulation‐based dynamic traffic assignment model DynusT (Dynamic Urban Systems in Transportation) is utilized to evaluate the system performance in response to different link‐additive toll schemes implemented on a highway in a real road transportation network. With the objective of minimizing the network‐wide average travel time, the simulation results show that implementing the optimal toll predicted by the surrogate model can benefit society in multiple ways. The travelers gain from the 2.5% reduction (0.45 minutes) of the average travel time. The total reduction in the time cost during the extended peak hours would be around US$65,000 for all the 570,000 network users assuming a US$15 per hour value of time. Meanwhile, the government benefits from the 20% increase of toll revenue compared to the current situation. Thus, applying the optimized pricing scheme in real world can be an encouraging policy option to enhance the performance of the transportation system in the study region.

Road tolling problem on a two-link parallel network based on the fundamental diagram

Routing and Congestion in Multi-Modal Transportation Networks

Surrogate‐Based Optimization of Expensive‐to‐Evaluate Objective for Optimal Highway Toll Charges in Transportation Network

Road pricing for congestion control with unknown demand and cost functions

Congestion Transportation Network and Optimal Environmental Link Tolls

Tracing Road Network Bottleneck By Data Driven Approach

Trial-And-Error Procedure for Implementing A Road-Pricing Scheme

On the Existence of System-Optimum–Inducing Tolling Functions for Transportation Networks

Second-best congestion pricing models for multiple time periods and multiple travelers

Parameter Estimation in Optimal Tolling for Traffic Networks Under the Markovian Traffic Equilibrium

Travel time function for basic link considering signal control in network traffic model

A Bi-Level Optimal Control Approach to Route Guidance Problem Considering Travelers' Stochastic Compliance

Highway toll allocation

A Continuous Transportation Network Design Problem with the Consideration of Road Congestion Charging

Comparison and optimization of cordon and area pricings for managing travel demand

Optimal Road Pricing for Both Traffic Efficiency and Safety

Optimizing Road Network Density Considering Automobile Traffic Efficiency: Theoretical Approach

Dynamic Tolling in Arc-based Traffic Assignment Models

Road System Optimization and Toll Design with Speed Limits

A Bilevel Multi-Objective Model for Sustainable Road Pricing Design on A Bimodal Transport Network

Trial-and-error Implementation of Marginal-Cost Pricing on Networks in the Absence of Demand Functions