Co-utile P2P ridesharing via decentralization and reputation management

Ridesharing has the potential to bring a wealth of benefits both to the actors directly involved in the shared trip (e.g., shared travel costs or access to high-occupancy vehicle facilities) and also to the society in general (e.g., reduced traffic congestion and CO2 emissions). However, even though ridesharing is based on a win-win collaboration and modern mobile communication technologies have significantly eased discovering and managing ride matches, the adoption of ridesharing has paradoxically decreased during the last years. In this respect, recent studies have highlighted how privacy concerns and the lack of trust among peers are crucial issues that hamper the success of ridesharing. In this paper, we tackle both of these issues by means of (i) a fully decentralized P2P ridesharing management network that avoids centralized ride-matching agencies (and hence private data compilation by such agencies); and (ii) an also decentralized reputation management protocol that brings trust among peers, even when they have not previously interacted. Our proposal rests on the recently proposed notion of co-utility (essentially, self-enforcing and mutually beneficial collaboration), which ensures that rational (even purely selfish) peers will find no incentives to deviate from the prescribed protocols. We have tested our system by using data gathered from real mobility traces of cabs in the San Francisco Bay area, and according to several metrics that quantify the degree of adoption of ridesharing and the ensuing individual and societal benefits.