As CO2 emission regulations increase, fleet owners increasingly consider the adoption of Electric Vehicle (EV) fleets in their business. The conventional Vehicle Routing Problem (VRP) aims to find a set of routes to reduce operational costs. However, route planning of EVs poses different challenges than that of Internal Combustion Engine Vehicles (ICEV). The Electric Vehicle Routing Problem (E-VRP) must take into consideration EV limitations such as short driving range, high charging time, poor charging infrastructure, and battery degradation. In this work, the E-VRP is formulated as a Prognostic Decision-Making problem. It considers customer time windows, partial midtour recharging operations, non-linear charging functions, and limited Charge Station (CS) capacities. Besides, battery State of Health (SOH) policies are included in the E-VRP to prevent early degradation of EV batteries. An optimization problem is formulated with the above considerations, when each EV has a set of costumers assigned, which is solved by a Genetic Algorithm (GA) approach. This GA has been suitably designed to decide the order of customers to visit, when and how much to recharge, and when to begin the operation. A simulation study is conducted to test GA performance with fleets and networks of different sizes. Results show that E-VRP effectively enables operation of the fleet, satisfying all operational constraints.
Reducing Waiting Times at Charging Stations with Adaptive Electric Vehicle Route Planning
