Scholarlink

<p>The facility layout problem (FLP) has been a prominent subject of research due to its practical significance and computational complexity. In the course of our study, we explored widely used benchmark problems and observed significant discrepancies in the reported outcomes for one particular instance, AB20. A prior study reported exceptionally good results on this problem, prompting further investigation. Our analysis revealed that, unlike other benchmarks, AB20 exhibits considerable variation in results due to inconsistencies in the underlying dataset, despite being referenced by the same name across multiple studies. This lack of standardization undermines the comparability of results and may lead to misleading conclusions. In this work, we systematically review previous studies that employed the AB20 benchmark and evaluate the consistency of the datasets used. Based on our findings, we propose guidelines to promote fair and reproducible benchmarking practices in FLP research. The study highlights the need for greater transparency and standardized datasets to ensure valid comparisons across future work.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Ensuring comparable experiment conditions in UA-FLP: The AB20 benchmark revisited

  • 期刊论文

  • 发表日期：

    2026年5月4日

查看全文
  • Junjae Chae,
  • Dongjin Jeong

摘要

The facility layout problem (FLP) has been a prominent subject of research due to its practical significance and computational complexity. In the course of our study, we explored widely used benchmark problems and observed significant discrepancies in the reported outcomes for one particular instance, AB20. A prior study reported exceptionally good results on this problem, prompting further investigation. Our analysis revealed that, unlike other benchmarks, AB20 exhibits considerable variation in results due to inconsistencies in the underlying dataset, despite being referenced by the same name across multiple studies. This lack of standardization undermines the comparability of results and may lead to misleading conclusions. In this work, we systematically review previous studies that employed the AB20 benchmark and evaluate the consistency of the datasets used. Based on our findings, we propose guidelines to promote fair and reproducible benchmarking practices in FLP research. The study highlights the need for greater transparency and standardized datasets to ensure valid comparisons across future work.