Evaluating Driver Strategies in Motorcycle-Based Ride-Hailing Services: An Agent-Based Modeling and Simulation Approach

The emergence of ride-hailing services has transformed urban transportation worldwide. These platforms enable users to book vehicles through smartphone applications and have been widely adopted across various regions. However, their rapid growth poses challenges related to service performance, environmental sustainability, and driver well-being. In Indonesia, motorcycle-based ride-hailing (MBRH) dominates urban transport due to its ability to navigate traffic congestion. This study employs an Agent-Based Modeling and Simulation (ABMS) approach integrated with Geographic Information Systems (GIS) to model interactions between customers and drivers within a spatial network replicating real-world conditions. The simulations evaluate various driver strategies, including prioritizing short-or long-distance rides, moving during idle times, and adopting a reactive strategy to maximize income. Key performance metrics such as passenger waiting time, dispatch distance, assigned-to-pick-up ratio, total distance traveled, order pick-up, and driver income assess platform performance and its impacts. A case study was conducted in East Surabaya, Indonesia. The findings illustrate how various driver strategies influence service outcomes. For instance, the short-distance strategy reduces passenger waiting times but limits driver productivity, whereas the long-distance strategy yields higher net income, especially during high-demand periods. There are trade-offs between service continuity and environmental impact, as random cruising modes improve assignment ratios but increase emissions due to longer travel distances. This research offers actionable insights for optimizing ride-hailing platforms and balancing benefits for drivers, customers, companies, and policymakers. It highlights the importance of sustainable strategies to enhance platform efficiency while promoting equitable growth within the ride-hailing ecosystem.