Solving Electric Vehicle Routing Problems using an Adaptive Large Neighbourhood Search

The Adaptive Large Neighbourhood Search of Pisinger & Ropke (2007) is able to solve various types of vehicle routing problems. Due to recent developments regarding global warming awareness and the usage of electric vehicles, the extension of this heuristic to electric vehicle routing problems is of interest. Therefore, this Adaptive Large Neighbourhood Search is implemented to test the replicability of the model, where after alterations to the electrical vehicle routing problem are made. In the Adaptive Large Neighbourhood Search, a solution is destroyed and subsequently repaired using destroy and repair heuristics that are chosen at random based on their historical performance. Pisinger & Ropke (2007) have proven the good performance of the model relative to other heuristics using the same data sets. Replicating the model did however show the dependency of the model on the parameter setting resulting in difficulties in reducing the number of required vehicles. The extended model for the electric vehicle routing model has shown great potential as well as the addition of extra destroy heuristics. The recommendation for future research is therefore to improve the model as to improve the vehicle minimisation stage and the additional destroy heuristics suggested in this paper.