A large neighbourhood based heuristic for two-echelon routing problems

In this paper we address an optimisation problem arising from city logistics. The focus lies on a two-level transportation system to deliver goods to customers within densely populated areas. The optimisation problem called the Two-Echelon Vehicle Routing Problem seeks to produce vehicle itineraries to deliver goods to customers with transit through intermediate facilities. A local-search metaheuristic based on the principle of destroy and repair of a Large Neighbourhood Search is developed and implemented to find high quality solutions within limited computing time. For future reference we resolve confusion with inconsistent versions of benchmark instances by explaining their differences and provide all of them online. The proposed algorithm is tested with those instances. It is able to find the currently best known solutions or better ones for 95% of the benchmark instances. The computational experiments show that this simple method achieves excellent solutions for the problem within short computing times.