Dynamic Carpool Matching for Employees: Leveraging Telematics and User Preferences

This paper presents a carpool matching framework designed to optimize the formation of efficient and feasible carpool groups. The framework incorporates various factors, including schedule compatibility, geographic proximity and user preferences such as cost savings, safety, and eco-friendliness, along with telematics data. Our approach dynamically forms carpool groups while allowing users to assign weights to specific criteria, ensuring that individual preferences are reflected in the final groupings. The experiments evaluate the framework’s performance across multiple dimensions: computational efficiency, scalability, and group quality. A synthetic dataset was generated to simulate urban commuting scenarios, incorporating employee home and work locations, work schedules, preferences, vehicle capacities, and driving behaviors. Execution time analysis demonstrates that the framework scales well, with acceptable computation times even for large datasets. We also assess the framework’s ability to form feasible carpool groups that meet logistical constraints and align with user preferences. Results indicate that the proposed framework significantly outperforms baseline methods, such as random matching and geographic proximity-based matching, in terms of feasibility rate, matching rate, and user satisfaction. This study demonstrates the potential of the carpool matching framework to support user-satisfying and sustainable carpooling solutions in large urban environments, while also providing insights into optimization areas for future work.