A practical framework for the electric vehicle routing problem

Sammanfattning: The routing of a delivery fleet is a classical optimization problem, known as the vehicle routing problem (VRP), which can heavily impact the quality of a logistic distribution process. Historically, the VRP formulation has mainly included internal combustion engine vehicles (ICEVs). However, due to their reduced environmental impact, the inclusion of electric vehicles (EVs) has become more popular. The inclusion requires accounting for a shorter driving range and limited infrastructure support. This thesis presents a framework for solving several practical extensions of the electric vehicle routing problem (E-VRP). Previously presented solvers usually target specific problem variants, optimize based on predetermined objectives, and display a general lack of discussion on their practical applicability. To counteract these shortcomings, the implemented framework allows for customizable objective functions, is capable of solving a wide range of practically relevant extensions, and provides an interface for investigating the properties of the proposed solutions. Examples of subjects treated are partial recharging, time-dependent variables, and dynamic settings. Solutions to real-world settings, modeled using the API of Iternio Planning AB, are demonstrated and the implemented solver shows promising results on a wide range of tested problem instances.