Improving off-line planning for vehicle routing problems in a time-dependent setting

In this research we evaluate how different vehicle routing strategies, that incorporate time-dependency on different levels, perform. In order to conduct this comparison, we construct several anonymized vehicle routing problem instances in Germany and use time-dependent travel time data from Bing's distance matrix API to evaluate those strategies. We show that by incorporating time-dependencies, delays at customers can be reduced by more than 60% with almost no increase in route duration. The results have been validated with Google's distance matrix API.