MARL-Based Pricing Strategy via Mutual Attention for MoD Systems with Ridesharing and Repositioning

The pricing strategy plays a vital role in increasing the revenue of mobility on-demand (MoD) systems by balancing supply-demand relationship across urban zones. MoD systems often incorporate both ridesharing and vehicle repositioning to maintain this balance, ultimately improving revenue and order completion rate. However, many existing pricing strategies overlook the effects when the price differences across zones meet ridesharing and repositioning, leading to supply-demand mismatch and reduced revenue. To tackle this problem, this paper presents a multi-agent reinforcement learning (MARL) based pricing strategy, named MAP, which employs a mutual attention mechanism to effectively account for price differences with ridesharing and repositioning. We transform the pricing problem as a MARL model that aiming to maximizing total revenue. These agents are tasked with making about order fare with ridesharing, vehicle income with repositioning for each zone. Pricing strategy decision-making is guided by mutual information theory and further enhanced by an attention mechanism to estimate how pricing variations impact adjacent zones. Simulation studies utilizing real-world datasets are performed to illustrate the advantages of MAP compared to existing benchmarks.