Scholarlink

<p>This research paper introduces MOIVYA (Multiobjective Ivy Algorithm) to solve multiobjective optimization problems based on inspiration drawn from ivy plant growth patterns. The ivy’s natural tendency to choose the nearest strong neighboring plant shapes is MOIVYA’s fundamental mechanism for solution enhancement and refinement. In this work, the traditional IVYA is improved by implementing elite non-dominated sorting, which enables accurate detection of Pareto optimal solutions during both offspring creation and selection. The efficient selection strategy of the algorithm uses a crowding distance-based approach to both improve convergence and maintain diversity among potential solutions. The evaluation of MOIVYA depends on three main metrics—hypervolume (HV), inverted generational distance (IGD), and computational runtime. The WFG test suite evaluation shows that MOIVYA provides superior performance compared to five other multiobjective optimization algorithms by reaching 8.5% better HV values alongside 12.3% lower IGD compared to the second-best algorithm. On tests with real-world optimization problems, too, MOIVYA is the best solution because it delivered better results than other algorithms while reducing HV by 15.7% together with a 10.4% decrease in IGD value. The MATLAB code is available at <a href="https://github.com/kanak02/MOIVYA">https://github.com/kanak02/MOIVYA</a>.</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 Novel Multi-objective Ivy Algorithm for Multi-criteria Engineering Design

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月21日

查看全文
  • Mohammad Aljaidi,
  • Amjad A. Alsuwaylimi,
  • Pradeep Jangir,
  • Mohammad Khishe

摘要

This research paper introduces MOIVYA (Multiobjective Ivy Algorithm) to solve multiobjective optimization problems based on inspiration drawn from ivy plant growth patterns. The ivy’s natural tendency to choose the nearest strong neighboring plant shapes is MOIVYA’s fundamental mechanism for solution enhancement and refinement. In this work, the traditional IVYA is improved by implementing elite non-dominated sorting, which enables accurate detection of Pareto optimal solutions during both offspring creation and selection. The efficient selection strategy of the algorithm uses a crowding distance-based approach to both improve convergence and maintain diversity among potential solutions. The evaluation of MOIVYA depends on three main metrics—hypervolume (HV), inverted generational distance (IGD), and computational runtime. The WFG test suite evaluation shows that MOIVYA provides superior performance compared to five other multiobjective optimization algorithms by reaching 8.5% better HV values alongside 12.3% lower IGD compared to the second-best algorithm. On tests with real-world optimization problems, too, MOIVYA is the best solution because it delivered better results than other algorithms while reducing HV by 15.7% together with a 10.4% decrease in IGD value. The MATLAB code is available at https://github.com/kanak02/MOIVYA.