A Dynamic Cordon Pricing Scheme Combining Macroscopic and Agent-Based Traffic Models

Congestion pricing is considered an effective management policy to reduce traffic congestion in transportation networks. This paper combines a macroscopic modeling of traffic congestion in urban networks with an agent-based simulator to study congestion pricing schemes in urban network. The macroscopic model, which has been tested with real data in previous studies, represents an accurate and robust approach to model the dynamics of congestion. The agent-based simulator can represent the complexity of travel behavior in terms of departure time choice and heterogeneous users. Traditional traffic simulators (including car-following, lane-changing and route choice models) consider traffic demand as input, i.e. inelastic to level of congestion conditions. On the other hand, most of traditional congestion pricing models, utilize a network supply curve which is not consistent with the physics of traffic and the dynamics of congestion, as they are sensitive to demand fluctuations and non-stationary conditions. Also, many of the existing pricing models are assuming deterministic and homogeneous population characteristics. In this paper, the authors first demonstrate by case studies in Zurich urban road network, that the output of a multi-agent based simulator, is consistent with the physics of traffic flow dynamics, as expressed by a macroscopic fundamental diagram (MFD). They then apply a dynamic cordon-based congestion pricing scheme, in which tolls are controlled by an MFD, and investigate the effectiveness of the proposed pricing scheme. Results show that by applying such a congestion pricing, (i) the savings of travel time at both aggregated and disaggregated level outweighs the costs, (ii) the congestion inside cordon is eased while no congestion is shifted to outside cordon, (iii) trips with purpose of going-to-work are less influenced by the charged toll than trips with purpose of going-for-leisure. Future work can apply the same methodology to other network-based pricing schemes. Equity issues can be investigated more carefully, if provided with data such as income of agents. Time-dependent or value-of-time-dependent pricing schemes then can also be determined.