Iterated local search for the CVRP with sequence-based pallet loading and axle weight restrictions

Abstract

An Iterated Local Search method (ILS) for the CVRP with sequence-based pallet loading and axle weight restrictions is presented. The problem considers the distribution of europallets to various locations and takes sequence-based loading as well as axle weight limits into account. Sequence-based loading ensures that no consignment is placed in such a way that it blocks the removal of items to be delivered earlier on the route. Pallets are placed in two horizontal rows inside the vehicle and cannot be vertically stacked. According to a survey among Belgian logistics service providers, axle weight limits impose a challenge for transportation companies since they face high fines when violating these limits. A metaheuristic as well as a matheuristic are developed for two variants of the problem. In the first problem type, pallets are packed dense, which means that there may be no gap between two consecutive pallets inside a vehicle. Dense packing makes it easier for the driver to secure the cargo than when pallets are spread over the vehicle. In the second problem type, gaps between pallets of consecutive customers are allowed. A single customer sequence therefore may result in different possible packing schemes. In this case, the feasibility check in terms of axle weight limits is solved exactly with Cplex. The resulting solution method is a matheuristic since the routing problem is solved heuristically while the loading feasibility check is computed exactly. The structure of the ILS as well as results will be presented.