A new iterative-doubling Greedy–Lookahead algorithm for the single container loading problem
► We developed a new Greedy–Lookahead tree search algorithm for the SCLP. ► We showed that selecting spaces is as important as selecting blocks of boxes. ► We proposed a simpler and more effective evaluation heuristic for selecting spaces. ► We considered the SCLP both with and without full support...
Gespeichert in:
Veröffentlicht in: | European journal of operational research 2012-11, Vol.222 (3), p.408-417 |
---|---|
Hauptverfasser: | , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | ► We developed a new Greedy–Lookahead tree search algorithm for the SCLP. ► We showed that selecting spaces is as important as selecting blocks of boxes. ► We proposed a simpler and more effective evaluation heuristic for selecting spaces. ► We considered the SCLP both with and without full support constraint. ► We outperformed all existing approaches based on 1600 benchmark instances.
The aim of the Single Container Loading Problem (SCLP) is to pack three-dimensional boxes into a three-dimensional container so as to maximize the volume utilization of the container. We propose a new block building approach that constructs packings by placing one block (of boxes) at a time until no more boxes can be loaded. The key to obtaining high quality solutions is to select the right block to place into the right free space cuboid (or residual space) in the container. We propose a new heuristic for evaluating the fitness of residual spaces, and use a tree search to decide the best residual space-block pair at each step. The resultant algorithm outperforms the best known algorithms based on the 1600 commonly used benchmark instances even when given fewer computational resources. We also adapted our approach to address the full support constraint. The computational results for the full support support variant on the 1600 instances similarly show a significant improvement over existing techniques even when given substantially fewer computational resources. |
---|---|
ISSN: | 0377-2217 1872-6860 |
DOI: | 10.1016/j.ejor.2012.04.036 |