The Electric Vehicle Route Planning Problem with Energy Consumption Uncertainty

Electric freight vehicles (EVs) are a sustainable alternative to conventional internal combustion freight vehicles. The driving autonomy of EVs is a fundamental component in the planning of EV routes for goods distribution. In this respect, a complicating factor lies in the fact that EVs' energy consumption is subject to a great deal of uncertainty, which is due to a number of endogenous and exogenous factors. Ignoring such uncertainties in the planning of EV routes may lead a vehicle to run out of energy, which-given the scarcity of recharging stations-may have dire effects. Thus, to foster a widespread use of EVs, we need to adopt new routing strategies that explicitly account for energy consumption uncertainty. In this paper, we propose a new two-stage stochastic programming formulation for the single electric vehicle routing problem with stochastic energy consumption. Furthermore, we develop a decomposition algorithm for this problem. We provide an illustrative example showing the added value of incorporating uncertainty in the route planning process. We perform a variety of computational experiments and show that our decomposition algorithm is capable of efficiently solving instances with 20 customers and 30 scenarios.