Altruistic Ride Sharing: A Community-Driven Approach to Short-Distance Mobility

TL;DR

The paper tackles urban mobility inefficiencies by proposing Altruistic Ride-Sharing (ARS), a decentralized, peer-to-peer framework where participants alternate between driver and rider roles based on altruism points rather than monetary incentives. ARS integrates a multi-agent reinforcement learning backbone (MADDPG), a game-theoretic fairness mechanism, and a biologically inspired population model, evaluated on real NYC TLC data to demonstrate reductions in travel distance, emissions, and congestion while boosting vehicle utilization and participation equity. The results show that ARS can outperform both no-sharing and optimization-based baselines, offering a scalable, community-driven alternative that aligns individual actions with urban sustainability goals. The work also outlines a trust and verification framework to support safe deployment and suggests future work on non-cooperative behavior and dynamic road topologies.

Abstract

Urban mobility faces persistent challenges of congestion and fuel consumption, specifically when people choose a private, point-to-point commute option. Profit-driven ride-sharing platforms prioritize revenue over fairness and sustainability. This paper introduces Altruistic Ride-Sharing (ARS), a decentralized, peer-to-peer mobility framework where participants alternate between driver and rider roles based on altruism points rather than monetary incentives. The system integrates multi-agent reinforcement learning (MADDPG) for dynamic ride-matching, game-theoretic equilibrium guarantees for fairness, and a population model to sustain long-term balance. Using real-world New York City taxi data, we demonstrate that ARS reduces travel distance and emissions, increases vehicle utilization, and promotes equitable participation compared to both no-sharing and optimization-based baselines. These results establish ARS as a scalable, community-driven alternative to conventional ride-sharing, aligning individual behavior with collective urban sustainability goals.

Figures (14)

• Figure 1: An overview of the daily cycle in the Altruistic Ride-Sharing (ARS) framework. Each day, community members opt into the system. Based on their altruism score, the system recommends a role (driver or rider). Users with scores above a minimum threshold can then make their final role choice. Drivers receive real-time rider pickup recommendations from the distributed routing engine. At the end of the day, all trips are evaluated to adjust user altruism scores. This creates a feedback loop where updated scores influence future role recommendations, and performance data helps refine the MARL model.
• Figure 2: Sankey Diagram for Role Assignment Logic. Participants with low altruism points ($\le0.2$) do not have the privilege to become a rider. Those with sufficient altruism points ($>0.2$) choose to become a rider/driver based in proportion with their normalized scores (green). Such participants can also show erratic behavior by randomly choosing between driver and rider.
• Figure 3: The Multi-Agent Reinforcement Learning (MARL) interaction loop in the Altruistic Ride-Sharing (ARS) system. An agent, trained via the MADDPG algorithm, receives its state observation (spatial coordinates, role, and local perception field) from the dynamic environment. It then takes a cooperative action (selecting or rejecting a rider) to maximize a multi-objective reward based on minimizing detours and increasing altruism.
• Figure 4: This flow diagram illustrates the agent "Birth-Death" process. New, unenrolled agents can join the system based on factors like network effects and system reputation. Once active, an agent's altruism score determines their probability of dropping out. Agents who have dropped out may later rejoin, a decision also influenced by their altruism score.
• Figure 5: The 15x15 grid overlay on the selected Manhattan corridor, which forms the simulation environment. The points represent the origins and destinations of trips sampled from the New York City TLC Yellow Taxi dataset to model realistic urban travel patterns. The figure also illustrates an example trip, showing a driver's path to accommodate riders.