Regulation of online hailing car cleanliness via heterogeneous evolutionary game with passenger selection

Abstract Despite the rapid growth of the online taxi industry, a critical component of Mobility as a Service (MaaS), the impact of vehicle cleanliness on service quality has been insufficiently explored. To bridge this gap, this paper employs heterogeneous evolutionary game theory combined with an agent-based model to simulate the interactions among drivers, passengers, and platforms with varying strategies toward cleanliness and complaint mechanisms. We examine three selection mechanisms: random, reputation-based, and sensor-based to assess their effects on promoting driver compliance with cleanliness standards and enhancing passenger satisfaction. Our findings reveal that reputation-based and sensor-based selection mechanisms can improve service quality and reduce passenger complaints compared to random selection. The reputation-based mechanism, while effective, tends to decrease drivers’ average payoffs, possibly affecting their long-term cooperation with platforms. In contrast, the sensor-based mechanism provides a balance by maintaining high service quality without substantially reducing driver income, suggesting a sustainable approach for platform policies. In addition, high rates of dirty drivers correlate with persistent noncompliance and elevated levels of passenger complaints, highlighting the need for enhanced complaint mechanisms and transparent monitoring systems. These insights underscore digital transportation platforms’ need to integrate advanced monitoring technologies and strategic passenger feedback systems to sustain high-quality service standards in the evolving ride-hailing market.