Three-axes rotation algorithm for the relaxed 3L-CVRP

Abstract

The purpose of this work is to present a developed three-axes rotation algorithm to improve the solving methodology for the relaxed 3L-CVRP (Three-Dimensional Capacitated Vehicle Routing Problem). Although there are reported works on solving approaches for the relaxed 3L-CVRP that consider product rotation to optimize load capacity, rotation on the three axes has not been thoroughly studied. In this aspect, the present work explicitly explores the three-axes rotation and its impact on load capacity optimization. In order to improve the relaxed 3L-CVRP problem, a two-phase solution was developed. The first phase consists of finding the solution for the CVRP problem, using a demand previously obtained with a heuristic developed to convert the 3L-CVRP demand into CVRP demand. The second phase is to obtain the loading of the vehicle using a heuristic developed to load the items using rules to obtain the rotation of the items. The proposed approach was able to improve the load assignment in 48.1% of well-known 3L-CVRP instances when compared to similar approaches on the relaxed 3L-CVRP. The outcomes of this research can be applied to transportation problems where package rotation on the z-axis is an option, and there are not fragile items to load in the vehicles.