Optimization of electric vehicle sharing network planning

Car sharing allows individuals to borrow cars from a designated area, for a rental rate charged by the time or the distance driven. Car sharing has social and environmental benefits such as the accessibility to more goods and services as well as the reduction of urban congestion and vehicle-kilometres travelled. Introducing electric vehicles (EVs) for car sharing may help reduce pollution and support the global effort to build an emission-free environment. In the UAE, it is important to promote the use of EVs to meet the objectives of the strategic initiative “UAE’s Net-Zero-2050”. Developing an EV sharing network may stimulate EV demand by providing easily accessible vehicles and charging stations at key locations to encourage drivers to try such new technologies. However, the design of an EV sharing network is a complex optimization problem that may involve thousands of possible locations for sharing/charging stations and millions of customer trip requests. Moreover, building sharing stations and chargers and acquiring a large EV fleet represent a huge investment. Therefore, decision-makers need appropriate decision-aid tools to optimally design EV sharing networks. In this project, we intend to develop such a decision-aid tool by relying on operations research and mathematical programming. We will focus on optimizing EV sharing station and charger locations as well as fleet size and deployment. We intend to improve the practical relevancy of our optimization model by considering two important features that have received little attention in the literature: the interrelations between strategic, tactical, and operational decisions and the uncertainty of several key input parameters. Thanks to the incorporation of these two features, our tool should provide governments or private organizations with relevant and practical strategic plans. To solve the resulting complex problem, we will develop innovative solution methods to obtain optimal or near-optimal solutions in reasonable computation times.