Large multiple neighborhood search for the soft-clustered vehicle-routing problem

The soft-clustered vehicle-routing problem (SoftCluVRP) is a variant of the classical capacitated problem. Customers are partitioned into clusters and all customers same cluster must be served by vehicle. In this paper, we present large multiple neighborhood search for SoftCluVRP. We design analyze destroy repair operators as well two post-optimization components, which both based on variable descent. first allows inter-route exchanges complete clusters, while second searches intra-route improvements combining neighborhoods (2-opt, Or-opt, double-bridge) Balas-Simonetti neighborhood. Computational experiments show that our algorithm clearly outperforms only existing heuristic approach from literature. By solving benchmark instances, provide 130 new best solutions 220 medium-sized instances with up to 483 prove 12 them optimal. Furthermore, describe straightforward adaption clustered problem,