Operations of a taxi fleet for advance reservations using electric vehicles and charging stations
This research investigated the operations of a taxi fleet that uses electric vehicles to accommodate only those trips for which advance reservations are made. In such operations, a dispatch system chains multiple trips to form a route and offers the route to a taxi driver. This unique problem is known as the Singapore taxi advance reservation with electric vehicles problem. Because electric vehicles have a limited running time before recharging is needed, the problem of minimizing the number of taxis used is formulated as a customized paired pickup and delivery problem with a time window and a charging station. A two-phase heuristic approach was proposed for solving this problem. Numerical experiments compared three solutions (nearest neighbor, sweep, and earliest time insertion heuristics), the effect of various battery recharging plans, and the effect of various numbers of charging stations for a simulated central business district network in Singapore. The results showed that (a) the earliest time insertion heuristic produced the best initial solutions; (b) as the maximum running time before recharging increased, the size of the fleet used decreased and there were fewer visits to the charging station, longer running times per taxi, and higher average revenues per driver; and (c) changing the number of charging stations had little effect on the system's performance measures.