École Nationale Supérieure de
Techniques Avancées

Abstract

Environmental concerns drive the need to rethink many aspects of the industry, and in particular the transportation sector. With the emergence of new vehicle technologies and environmental policies, fleet operators face the challenging transition problem of choosing their fleet composition and daily activities, at best financial cost and lowest carbon footprint. With the distant aim of creating a tool for decision support for fleet operators, this work tackles a particular example of an electric vehicle routing problem, with time windows and pick-up and delivery activities. We propose a formulation of the problem as a MILP, with a particular node-multiplication model, and a heuristic approach. This heuristic approach revolves around a labeling algorithm creating elementary routes and an oracle, forming a natural greedy algorithm.