Scholarlink

<p>The Quadratic Assignment Problem (QAP), a fundamental NP-hard combinatorial optimization challenge, requires efficient algorithms to navigate complex search spaces. This study introduces a novel hybrid metaheuristic that integrates the Sailfish Optimizer (SFO), inspired by sailfish hunting strategies, with the 2-opt local search to tackle the QAP. Evaluated on instances from the Quadratic Assignment Problem Library (QAPLIB), the proposed algorithm achieves the best-known solution for 50 of 53 selected QAPLIB benchmark instances, with solution deviations ranging from 0 to 0.19% and execution time varying from 0.8 to 14.6&#xa0;s for small and medium instances (n ≤ 64) and up to 90&#xa0;s for the largest instance (tho150). A comparative analysis demonstrates its superior performance over eight state-of-the-art metaheuristics, including Genetic Hybrid and Parallel Biogeography-Based Optimization with Tabu Search, as evidenced by Wilcoxon signed-rank and analysis-of-variance tests. The key contributions of this work include a hybrid framework that effectively balances exploration and exploitation, tailored enhancements to the SFO and 2-opt methods for the QAP, comprehensive evaluations demonstrating scalability, and behavioral analysis that provides valuable insights into metaheuristic design.</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

SFO-QAP: a novel hybrid sailfish optimizer with 2-opt for efficient quadratic assignment problem optimization

  • 期刊论文

  • 发表日期：

    2026年2月23日

查看全文
  • Khaoula Cherrat,
  • Mourad Mzili,
  • Toufik Mzili,
  • Nebojsa bacanin

摘要

The Quadratic Assignment Problem (QAP), a fundamental NP-hard combinatorial optimization challenge, requires efficient algorithms to navigate complex search spaces. This study introduces a novel hybrid metaheuristic that integrates the Sailfish Optimizer (SFO), inspired by sailfish hunting strategies, with the 2-opt local search to tackle the QAP. Evaluated on instances from the Quadratic Assignment Problem Library (QAPLIB), the proposed algorithm achieves the best-known solution for 50 of 53 selected QAPLIB benchmark instances, with solution deviations ranging from 0 to 0.19% and execution time varying from 0.8 to 14.6 s for small and medium instances (n ≤ 64) and up to 90 s for the largest instance (tho150). A comparative analysis demonstrates its superior performance over eight state-of-the-art metaheuristics, including Genetic Hybrid and Parallel Biogeography-Based Optimization with Tabu Search, as evidenced by Wilcoxon signed-rank and analysis-of-variance tests. The key contributions of this work include a hybrid framework that effectively balances exploration and exploitation, tailored enhancements to the SFO and 2-opt methods for the QAP, comprehensive evaluations demonstrating scalability, and behavioral analysis that provides valuable insights into metaheuristic design.