Large neighborhood search for the vehicle routing problem with two-dimensional loading constraints

In this paper we investigate the use of the large neighborhood search for solving the vehicle routing problem with two-dimensional loading constraints, an NP-hard combinatorial optimization problem. Such a problem may be viewed as the combination of two complementary well-known problems: two-dimensional bin-packing and capacitated vehicle routing. The proposed method considers a two-phase solution procedure: a first phase generates a placement order and applies a bottom-left strategy to this order for packing the items in the vehicle and, a second phase is applied for reaching a feasible routing. Both phases are combined by using a diversification strategy. The proposed method has been evaluated on Gendreau's benchmark instances. The obtained results are compared to those reached by the best methods taken from the literature. Encouraging results have been obtained.