Dynamic Pricing Analysis under Demand-Supply Equilibrium of Autonomous-Mobility-on-Demand Services

Autonomous Mobility-on Demand (AMoD) combines self-driving and Mobility-on-Demand (MoD) services, allowing passengers to enjoy the last mile service. Due to the success of Uber, Lift, and other ride-sourcing companies, previous research has discussed pricing strategies in the ride-sourcing market. This study combines ridesharing and dynamic pricing strategy in the AMoD system, building models to maximize social welfare. Shared Autonomous Vehicles (SAV) ridesharing and dispatching models are constructed, and numerical experiments are conducted on a real road network using different factors. Linear regression models based on the simulation data from the dispatching model are established to predict the average waiting time and meeting rate. The regression models are applied to the ride-sourcing market model to conduct dynamic pricing experiments. We use approximate dynamic programming (ADP) to solve the dynamic pricing multiplier for each time interval. The experimental results show that ridesharing can improve the service rate of rides, and the dynamic pricing strategy achieves higher social welfare by balancing supply and demand compared to the fixed pricing strategy.