A Tree Search-Based Heuristic for the Three-Dimensional Single Container Loading Problem

A Tree Search‑Based Heuristic for the Three‑Dimensional Single Container Loading Problem This repository contains all the research data. You can find a brief description of each file below. sclp_app_docker.tar.gz: Docker image (tarball) containing the SCLP application environment, ready for deployment and reproducible experiments. You can find how to run it in the README in 3d_sclp_source_code. lucasguesserts_masters_thesis.pdf: PDF version of Lucas Guesser’s master’s thesis, providing background, methodology, and results related to the research. Note that there were some change to the name of the heuristics from the dissertation to the paper published. 3d_sclp_source_code.zip: Source code of the heuristic proposed. This code is also available in https://github.com/lucasguesserts/sclp-heuristic araya_et_al_2017_source_code.zip: Original source code accompanying the 2017 Araya et al. study, used here for baseline comparisons. Araya et al. have their repository with the code: https://github.com/rilianx/Metasolver The authors have their fork of that repository. We hope ours is easier to compile: https://github.com/lucasguesserts/Metasolver experimental_results.zip: Archive of all raw experimental output files and analyses generated during the study. instance_generator_bischoff_and_ratcliff_1995.zip: A generator of the Bischoff & Ratcliff (1995) benchmark instances. instances_1995_bischoff_and_ratcliff_and_1999_davies_and_bischoff.zip: The benchmark instances from the 1995 Bischoff/Ratcliff and 1999 Davies/Bischoff datasets. In addition to the files in our file format, we also provide the files provided by other authors. lucasguesser_master_thesis.zip: Full archive of Lucas Guesser’s master thesis source files (LaTeX, figures, bibliography). It can also be found in https://www.repositorio.unicamp.br/Busca/Download?codigoArquivo=582099 packing-visualizer.zip: Web-based tool to visualize the packings produced by the proposed heuristic. This code is also available in https://github.com/lucasguesserts/packing-visualizer scripts.zip: Scripts (data preprocessing, post‑processing, and automation) used throughout the experiments. It contains: solve: a script that automates the whole process of executing the heuristic in parallel with many different configurations. tuning/: scripts and configurations used to run irace. figure_tools: scripts used to create the figures of the paper.面向三维单容器装载问题（Three-Dimensional Single Container Loading Problem）的基于树搜索的启发式算法 本仓库收录本研究的全部科研数据，下文将对各文件进行简要说明。 sclp_app_docker.tar.gz：包含SCLP应用环境的Docker镜像（tar压缩包格式），支持直接部署以开展可复现实验。其运行方法可参见3d_sclp_source_code目录下的README文档。 lucasguesserts_masters_thesis.pdf：Lucas Guesser硕士学位论文的PDF版本，涵盖本研究的背景、方法论与实验结果。需注意：本研究所用启发式算法的命名与学位论文发表时存在小幅调整。 3d_sclp_source_code.zip：本文所提启发式算法的源代码。该代码同时托管于GitHub仓库：https://github.com/lucasguesserts/sclp-heuristic araya_et_al_2017_source_code.zip：对应Araya等人2017年研究的原始源代码，本研究将其用作基准对照实验的实现。Araya等人的官方代码仓库为：https://github.com/rilianx/Metasolver。本研究使用了该仓库的分支版本，我们的分支版本更易于编译：https://github.com/lucasguesserts/Metasolver experimental_results.zip：本研究实验过程中产生的全部原始实验输出文件与分析结果归档包。 instance_generator_bischoff_and_ratcliff_1995.zip：Bischoff与Ratcliff（1995）基准测试实例生成器。 instances_1995_bischoff_and_ratcliff_and_1999_davies_and_bischoff.zip：收录1995年Bischoff/Ratcliff与1999年Davies/Bischoff数据集的基准测试实例文件。 除本研究自定义格式的文件外，本仓库同时提供其他作者发布的原始文件。 lucasguesser_master_thesis.zip：Lucas Guesser硕士学位论文的完整源文件归档包（包含LaTeX源码、插图与参考文献列表）。该文件亦可通过以下链接获取：https://www.repositorio.unicamp.br/Busca/Download?codigoArquivo=582099 packing-visualizer.zip：用于可视化本启发式算法生成的货物装载方案的Web工具。该代码同时托管于GitHub仓库：https://github.com/lucasguesserts/packing-visualizer scripts.zip：本研究实验全程所用的脚本文件（涵盖数据预处理、后处理与自动化执行脚本），具体包括： - solve：用于并行执行多配置下启发式算法的自动化脚本，覆盖完整执行流程 - tuning/：用于运行irace调参工具的脚本与配置文件 - figure_tools：用于生成论文配图的脚本文件