Scholarlink

<p>In this research, a highly effective population evolution-based multi-strategy improved lemur optimiser (MILO) algorithm is proposed to tackle the flaws of the existing lemur optimiser (LO) algorithm, including issues like slow convergence and an imbalanced strategy for exploration and exploitation. First, Chebyshev chaotic mapping is employed to heighten the diversity and optimisation efficiency of the initial population. Second, to avert the algorithm from converging to local optima, a multielite-guided differential population evolution strategy is proposed. Finally, the algorithm integrates longitudinal and transversal crossover strategies to enable comprehensive solution space exploration, thereby enhancing the convergence accuracy in later evolutionary stages. To evaluate the performance of MILO, simulation experiments are conducted using the CEC2005, the CEC2017, and the CEC2022 function set, demonstrating that MILO outperforms eleven well-established swarm intelligence optimisers, especially in terms of solution accuracy and stability. Additionally, compared with the original algorithm, the MILO improves the solution accuracy rate by 19.56%, 25.79%, and 45.73% in the optimisation of three typical engineering design problems. MILO can effectively solve complex multi-objective problems and holds potential for application in intricate real-world optimisation problems.</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

A multi-strategy improved lemur optimiser for global optimisation and engineering applications

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月28日

查看全文
  • Mingjia Li,
  • Dexin Sun,
  • Qianliang You,
  • Chengsheng Ji,
  • Jing Xu

摘要

In this research, a highly effective population evolution-based multi-strategy improved lemur optimiser (MILO) algorithm is proposed to tackle the flaws of the existing lemur optimiser (LO) algorithm, including issues like slow convergence and an imbalanced strategy for exploration and exploitation. First, Chebyshev chaotic mapping is employed to heighten the diversity and optimisation efficiency of the initial population. Second, to avert the algorithm from converging to local optima, a multielite-guided differential population evolution strategy is proposed. Finally, the algorithm integrates longitudinal and transversal crossover strategies to enable comprehensive solution space exploration, thereby enhancing the convergence accuracy in later evolutionary stages. To evaluate the performance of MILO, simulation experiments are conducted using the CEC2005, the CEC2017, and the CEC2022 function set, demonstrating that MILO outperforms eleven well-established swarm intelligence optimisers, especially in terms of solution accuracy and stability. Additionally, compared with the original algorithm, the MILO improves the solution accuracy rate by 19.56%, 25.79%, and 45.73% in the optimisation of three typical engineering design problems. MILO can effectively solve complex multi-objective problems and holds potential for application in intricate real-world optimisation problems.