Non-myopic dynamic routing of electric taxis with battery swapping stations

Electric vehicles (EVs) have a key aspect in reducing the greenhouse gas effect, maintenance, and energy expenditures of drivers. A type of integration of the electric vehicle (EV) charging infrastructure is emerging based on the premise of battery swapping. Drivers can exchange their empty batteries quickly with full batteries from any battery swapping station. The limited battery capacities of electric taxis require visiting the swapping stations during pickup and drop-off tours, which entails choosing the route more effectively to avoid customer delay. We propose the dynamic routing of electric taxis with a look-ahead policy using a Markov decision process (MDP) for assigning an electric taxi fleet to customers with the assumption of elastic demand. This is the first model that focuses on the design of a non-myopic routing of electric taxis that considers the limited battery capacity through serving customers, detours of the taxi drivers to battery swapping stations (BSS), and integration of customers delay and the system cost into a dynamic non-myopic pricing policy under the objective of maximizing social benefit. Using battery recharging locations and taxicab trip data in New York City, we showed an improvement in the average of social welfare, due to use of clean and smart taxi routes based on the proposed dynamic non-myopic routing policy by up to 8% compared to the routing problem without a look-ahead policy.