Evaluation of Dynamic Incentive Pricing for Congestion Management in Transit System: An Agent-Based Simulation

This paper analyzed the feasibility of a dynamic fare incentive strategy by characterizing commuters’ travel patterns and the extent of their flexibility in departure times using multi-source data. In the proposed fare incentive, commuters incur a surcharge during the central peak period and obtain a monetary reward during the shoulder peak period. The fare incentive determines central period location and length, the value of reward and the reward ratio which is the ratio of the number of trips with reward to total number of trips. We proposed an agent-based simulation to evaluate the performance of the fare incentive inclusive of outlier analysis, travel pattern recognition, and crowdedness interpretations. The simulation and evaluation are applied to a metropolitan transit system using smartcard and operation data. Results reveal the practicality of the proposed fare incentive to reduce the congestion by affecting commuter departure time distribution while keeping the flexibility interval unchanged.