An Iterated Tabu Search for the Vehicle Routing Problem with Multiple Compartments and Last-in-First-Out Unloading

We study a variant of the vehicle routing problem with multiple compartments in which compartment sizes are flexibly adjustable in discrete steps. Additionally, a last-in-first-out unloading policy has to be considered since rearranging of cargo on the load bed is prohibited. This problem occurs as a special case at a German food retailer that has to supply a large number of hypermarkets on a daily basis. Transported products are divided into several categories based on the transport temperature. Due to these different temperatures, it is expedient to use trucks with multiple compartments. Moreover, the retailer engages various forwarders for transportation. Forwarders are paid according to previously negotiated tariffs based on the different product categories. This leads to a cost function that differs a lot from the common distance minimization. We propose an iterated tabu search with two different perturbation mechanisms to solve this challenging vehicle routing problem. Various computational experiments with real data from the retailer are performed to assess the performance of our algorithm. Comparisons with results obtained by employing a mixed integer program and solving this with a commercial solver show that the iterated tabu search consistently produces high quality solutions.