Quality of Experience Oriented Eco-Friendly Taxi-Ride Sharing Recommendation Framework

Abstract

Carpooling services are increasingly popular due to their potential for reducing fares, shortening travel times, and increasing driver income. However, despite the efforts of many carpooling recommendation systems to balance the often conflicting objectives of drivers and passengers, there remains significant room for improvement in effectively managing these objectives. To address this challenge, a novel quality of experience-oriented, eco-friendly taxi-ride recommendation system called QE-Ride is proposed. This system optimizes taxi selection for passengers by evaluating factors such as time delay tolerance, preferred vehicle capacity, fare reduction preferences, tolerance for additional driving distance, as well as preferences for driver safety, eco-friendliness, and the past ratings of both drivers and co-travelers. For drivers, QE-Ride considers their preferred vehicle capacity, interest in profit maximization, and passengers’ past ratings. The effectiveness of QE-Ride was validated using GPS trace data from 10,357 taxicabs, demonstrating the system’s ability to align with both drivers’ and passengers’ incentives while accounting for real-time traffic conditions. The results indicate that QE-Ride outperforms existing systems in key areas, including reducing total mileage, lowering passenger fares, and increasing driver profits. By introducing a balanced approach that considers eco-friendliness, safety, and other critical factors, QE-Ride offers a promising enhancement to the overall carpooling experience, making it a standout tool in the domain of ride-sharing services.

Keywords

• Intelligent transportation system
• carpooling
• dynamic ride-sharing
• vehicle recommendation