MANAGEMENT OF A SHARED, AUTONOMOUS, ELECTRIC VEHICLE FLEET: CHARGING AND PRICING STRATEGIES

The nexus of autonomous vehicle (AV) and electric vehicle (EV) technologies has important potential impacts on our transportation systems, particularly in the case of shared-use vehicles. Incentivizing plug-in EV adoption and shared-vehicle use may be a key strategy for helping regions achieve air quality standards for ozone and particulate matter, and ultimately greenhouse gas emissions standards. Various energy pricing, vehicle design, vehicle pricing, power feedstock, and other factors all affect PEV adoption, use, and charging patterns. Such choices affect emissions totals, air quality, and climate change. At the same time, carsharing is emerging as an alternative mode that is more flexible than transit and less expensive than traditional private vehicle ownership. Carsharing has multiple environmental impacts: vehicle ownership and miles-traveled generally fall, along with embodied and operational energy and emissions, and infrastructure demands. The growth of EVs and carsharing are both hindered by technological and social factors. For battery-only EVs, a key challenge is “range anxiety” - a user’s concern for being stranded with a fully discharged battery and no reasonable recharge option (Bartlett 2012). Fortunately, EVs are a natural match for carsharing operations, since shared vehicles tend to be smaller and more fuel efficient than privately held vehicles (see, e.g., Meijkamp 1998, Martin and Shaheen 2011, Ryden and Morin 2005). Cutting-edge carsharing operators (CSOs) are already employing EVs in their fleets (such as Daimler’s Car2Go and BMW’s DriveNow operations), but the CSO’s having to relocate individual vehicles in one-way car sharing systems presents cost and profitability challenges. Level-4 (fully self-driving) AV technology can remove the barrier of manual vehicle relocation and provide a driver-free method for shared EVs to reach travelers’ origins and destinations, as well as charging stations. In a carsharing setting, a fleet of shared autonomous (battery-only) electric vehicles (SAEVs) can remove most, if not all, range-anxiety issues while automating the battery management and charging process. With rising EV sales and carsharing membership levels, around the world, and the recently popularity of on-demand transportation services through transportation network companies, it is possible to imagine a future travel system where AV technologies merges with these transportation trends in a SAEV fleet. Personal transportation will be as easy as hailing a SAEV through a computer or mobile device, waiting for the SAEV to arrive, and sitting comfortably while the SAEV drives to the destination. But can SAEVs be shared, self-charged, and right (battery-) sized for the trip lengths that travelers desire? This paper explores the management of a fleet of SAEVs in a discrete-time agent-based model based on the SAV framework introduced in Fagnant and Kockelman’s (2014) 10 mile by 10 mile simulation. This new regional (100 miles by 100 miles) and electrified simulation employs multiple charging strategies for the SAEVs and seeks optimal charging station locations in order to moderate un-occupied (empty-vehicle) travel distances to and from charging stations while maximizing CSO profits and/or meeting all demands with minimal delay. The system’s performance (e.g., average wait times, number of fleet vehicles required, un-occupied travel distances) is examined in relation to the number of charging stations, the electric range of the vehicles, and the type of charging system (and the corresponding length of charging time). Such analysis examines the various cost tradeoffs decisions for fleet operators. For example, the additional cost of Level 3 DC (fast charging) stations means shorter charging times, and translates to a smaller fleet than one that would be required with slower and less expensive Level 2 AC charging stations. In addition to charging location and scheduling decisions, the work examines pricing strategies to maintain vehicle balance (over space, relative to demand). An SAEV fleet operator seeks to maximize profit but must maintain some base level of service (e.g., share of trip requests served and average wait time). Using a block balancing technique as one of many methods to determine surcharges and credits at each trip’s origin and destination, the pricing strategies will incentivize travel to zones in need of more vehicles and penalize travel to zones with an excess of vehicles. A mode choice step allows each trip to be associated with a unique traveler (and his/her willingness to pay), and the pricing strategies correspondingly motivate trip patterns that benefit vehicle balance over spatial demands. Price deductions to reflect delays will also be considered, to incentivize choice of the SAEV mode. Preliminary results suggest that each SAEV can replace 6.5 privately owned vehicles in a 100 mile by 100 mile region, serving 96% of trips with an average wait time of 7.5 minutes. At the same time, the SAEV fleet is predicted to induce 20% empty-vehicle travel.