Optimization for dynamic ride-sharing: A review

Abstract

Dynamic ride-share systems aim to bring together travelers with similar itineraries and time schedules on short-notice. These systems may provide significant societal and environmental benefits by reducing the number of cars used for personal travel and improving the utilization of available seat capacity. Effective and efficient optimization technology that matches drivers and riders in real-time is one of the necessary components for a successful dynamic ride-share system. We systematically outline the optimization challenges that arise when developing technology to support ride-sharing and survey the related operations research models in the academic literature. We hope that this paper will encourage more research by the transportation science and logistics community in this exciting, emerging area of public transportation.

Introduction

Finite oil supplies, rising gas prices, traffic congestion, and environmental concerns have recently increased the interest in services that allow people to use personal automobiles more wisely. The demand for ride-sharing services, which aim to bring together travelers with similar itineraries and time schedules, has increased sharply in recent years (Saranow, 2006). Ride-share providers across the globe are offering online notice boards for potential carpoolers, whether for daily commutes or for one-time trips to festivals, concerts, or sports events. Some online services, such as Nuride, provide incentives like restaurant coupons, gift certificates, or retail sales discounts to participants. Ride-sharing has generated much interest, and recent media coverage can be found in the Wall Street Journal (Saranow, 2006), Newsweek (Levy, 2007), Business Week (Walters, 2007), ABC News (Bell, 2007), The NY Times (Wiedenkeller, 2008), among many others.

Private car occupancy rates (the number of travelers per vehicle trip) are relatively low; average car occupancies in Europe range from 1.8 for leisure trips to 1.1 for commuters (EEA, 2005). Similar occupancy rates are also found in the US (Santos et al., 2011). The large demand for automobile transportation at peak-hours together with low occupancies leads to traffic congestion in many urban areas. The annual cost of congestion in the US in terms of lost hours and wasted fuel was estimated to be $78 billion in 2007 (Schrank and Lomax, 2007). Private automobile usage is also the dominant transportation mode producing carbon dioxide emissions (Hensher, 2008). Vehicle emissions give rise to problems both on a local and global scale. Locally, the health effects of air pollution represent a serious problem in many of the most densely populated regions worldwide (Brunekreef and Holgate, 2002, Kunzli et al., 2000). Globally, carbon dioxide emissions are associated with climate change and global warming.

Effective usage of empty car seats by ride-sharing may represent an important opportunity to increase occupancy rates, and could substantially increase the efficiency of urban transportation systems, potentially reducing traffic congestion, fuel consumption, and pollution. Moreover, ride-sharing allows users to share car-related expenses, which can be substantial. While ride-sharing is not a new idea, recent technological advances should increase its popularity, as we will explain. Certainly, ride-sharing must be easy, safe, flexible, efficient and economical before it will be adopted more widely.

To broaden its appeal, ride-sharing must be able to compete with one of the greatest advantages of private car usage: immediate access to door-to-door transportation. Technological advances, both hardware and software, are key enablers. In the US, the number of smartphone subscribers using the mobile Internet has grown 45 percent since 2010 (Nielsen, 2011), with approximately 40% of the mobile subscribers using a smartphone in 2011 (Smith, 2011). Similar estimates apply to the leading European markets (comScore, 2011). The growing ubiquity of Internet-enabled mobile devices partially enables practical dynamic ride-sharing (Hartwig and Buchmann, 2007, Hartmann, 2008).

By dynamic ride-sharing, we refer to a system where an automated system made available by a ride-share provider matches up drivers and riders on very short notice or even en-route. Recent startups like Carticipate, EnergeticX, Avego, and Flinc offer dynamic ride-sharing applications that allow drivers with spare seats to connect to people wanting to share a ride. They provide applications that run on (location-aware) Internet-enabled mobile phones. To ease the fear of sharing a ride with a potential stranger, these services use reputation systems (see e.g., PickupPal) or can be linked with social network tools like Facebook (see e.g., GoLoco and Zimride).

The ability of a dynamic ride-share provider to successfully establish ride-shares on short notice depends on the characteristics of the environment in terms of participant geographic density, traffic patterns, and the available roadway and transit infrastructure. Hall and Qureshi (1997) analyze the likelihood that a person will be successful in finding a ride-match, given a pool size of potential ride matches. Using probabilistic analysis, they conclude that in theory ride-sharing is viable since a congested freeway corridor should offer sufficient potential ride-matches. The authors also observe that there are many obstacles, primarily in terms of communication, so that the chance of finding a ride match in practice may in fact be small. Fortunately, technological advances have greatly reduced this communication obstacle.

Although the enabling technology is available, ride-sharing success stories are still in short supply. The development of algorithmic approaches for optimally matching drivers and riders in real-time may only play a small role in the ultimate success of ride-sharing, but it is central to the concept. Therefore, we believe that the time is right for a systematic overview of the issues in ride-sharing and the relevant optimization models that support the matching of riders and drivers in real-time. Since the operations research community has only recently started to address the related optimization challenges, we not only focus on the literature that specifically considers ride-share optimization but also on optimization approaches in other areas of transportation that share similar features. By introducing and formally defining dynamic ride-sharing problems, and by illustrating and outlining the optimization challenges that arise when developing technology to support ride-sharing, we hope to encourage more research by the transportation science and logistics community in this exciting, emerging area of public transportation.

The remainder of the paper is structured as follows. In Section 2, we explain and characterize the dynamic ride-sharing concept and introduce several relevant planning issues that arise in this context. In Section 3, we present a more formal definition of the basic ride-sharing problem and its various variants and survey the available literature. In Section 4 we discuss dynamic ride-sharing problems. In Section 5, we present the multi-modal version of the ride-sharing problem. Finally, in Section 6, we summarize our main insights and discuss directions for future research. Throughout the paper we also provide an overview of related problems in passenger and freight transportation optimization.