Sharing the Shared Rides: Multi-Party Carpooling Supported Strategy-Proof Double Auctions.

This article focuses on the design of strategy-proof double auction mechanisms to support multi-party carpooling, where multiple riders with overlapping itineraries share rides from a single driver. It introduces the MASTER (Multi-party cArpooling SupporTed stratEgy-pRoof) family of mechanisms tailored for both scheduled and on-demand carpooling scenarios, addressing challenges such as personalized carpooling constraints, information asymmetry, and strategic bidding. The scheduled mechanisms (MASTERmS and MASTERMS) differ in how group bids are determined and include filtration processes to maintain individual rationality, while the on-demand mechanisms (MASTERO) incorporate frustration-based promotion to prioritize riders with longer waits and provide operational flexibility via adjustable promotion strength. Experimental results based on real-world data demonstrate that multi-party carpooling improves social welfare and car occupancy rates, with the proposed mechanisms outperforming existing academic and practical counterparts in allocation efficiency and service responsiveness, especially under varying market conditions and rider-driver ratios.