Scholarlink

<p>The multi-objective cuckoo search (MOCS) algorithm is an effective metaheuristic for solving complex optimization problems. Like many optimization algorithms, MOCS faces challenges with premature convergence and maintaining diversity in the solution space. To address this issue, we propose an upgraded MOCS algorithm that includes a migration operator that is activated upon early convergence detection. Migration operator aims to improve exploration and exploitation by reintroducing diversity when the algorithm stagnates. The proposed algorithm, termed MG-MOCS, is evaluated using standard benchmark functions, including IMOP and ZDT problems. MG-MOCS shows significant improvements in spacing and hypervolume metrics. Statistical tests such as the Friedman test confirm the superior performance of MG-MOCS compared to other state-of-the-art algorithms like NSGA-II, NSGA-III, SPEA2, MOGWO and MSKEA. Furthermore, MG-MOCS is applied to a real-world IoT task scheduling problem in a cloud-Fog computing environment, where it efficiently balances workload distribution across Cloud and Fog resources. Results demonstrate that the migration operator enhances MOCS’s ability to avoid local optima, resulting in better optimization performance.</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

Enhanced multi-objective cuckoo search with migration operator for benchmark optimization and IoT task scheduling in cloud-fog computing

  • 期刊论文

  • 发表日期：

    2025年6月13日

查看全文
  • Fatemeh BahraniPour,
  • Mohammad Farshi,
  • Sepehr Ebrahimi Mood

摘要

The multi-objective cuckoo search (MOCS) algorithm is an effective metaheuristic for solving complex optimization problems. Like many optimization algorithms, MOCS faces challenges with premature convergence and maintaining diversity in the solution space. To address this issue, we propose an upgraded MOCS algorithm that includes a migration operator that is activated upon early convergence detection. Migration operator aims to improve exploration and exploitation by reintroducing diversity when the algorithm stagnates. The proposed algorithm, termed MG-MOCS, is evaluated using standard benchmark functions, including IMOP and ZDT problems. MG-MOCS shows significant improvements in spacing and hypervolume metrics. Statistical tests such as the Friedman test confirm the superior performance of MG-MOCS compared to other state-of-the-art algorithms like NSGA-II, NSGA-III, SPEA2, MOGWO and MSKEA. Furthermore, MG-MOCS is applied to a real-world IoT task scheduling problem in a cloud-Fog computing environment, where it efficiently balances workload distribution across Cloud and Fog resources. Results demonstrate that the migration operator enhances MOCS’s ability to avoid local optima, resulting in better optimization performance.