Scholarlink

<p>This paper presents a new metaheuristic optimization technique, termed earthquake dynamics-based optimization (EDBO), inspired by the physical behavior of seismic waves as they travel through heterogeneous layers of the Earth. Global search performance is enhanced by EDBO’s effective maintenance of a strong balance between exploration and exploitation through the modeling of P-wave and S-wave propagation and the integration of the shadow zone concept. The algorithm finds optimal solutions for complex optimization problem through a detailed search-space exploration and employs a dynamic population regulation mechanism that simulates seismic activity to preserve diversity and prevent premature convergence. A single-objective framework is employed to ensure computational efficiency, clarity, and simplicity in model evaluation. Using the CEC 2021 and CEC 2024 benchmark suites, the proposed algorithm is evaluated and shows good accuracy, speed of convergence, and scalability when dealing with single-objective bound-constrained problems. With a total score of 94.668, EDBO places second in CEC 2021, and its efficiency is further supported by its runtime analysis. Statistical evaluations, including standard error analysis and ranking performance, validate its robustness. Additionally, the results of the Friedman and Wilcoxon rank-sum tests demonstrate that EDBO routinely performs better than well-known algorithms such as jellyfish search optimizer (jSOA), modified L-SHADE (mLSHADE-RL), linear success rate-based adaptive differential evolution (L-SRTDE), and reconstructed differential evolution (RDE), with statistically significant improvements (<i>p</i> &lt; 0.05) and achieving a mean Friedman rank of 4.09, EDBO outperform algorithms including ocean water current optimization (OWCO) and many more. These results validate EDBO’s robustness, adaptability, and superior performance across diverse benchmark functions.</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

Earthquake dynamics-based optimization: a transformational tool for single-objective bound-constrained problems

  • 期刊论文

  • 发表日期：

    2026年4月2日

查看全文
  • Siddhi Kumari Sharma,
  • Lavika Goel,
  • Namita Mittal

摘要

This paper presents a new metaheuristic optimization technique, termed earthquake dynamics-based optimization (EDBO), inspired by the physical behavior of seismic waves as they travel through heterogeneous layers of the Earth. Global search performance is enhanced by EDBO’s effective maintenance of a strong balance between exploration and exploitation through the modeling of P-wave and S-wave propagation and the integration of the shadow zone concept. The algorithm finds optimal solutions for complex optimization problem through a detailed search-space exploration and employs a dynamic population regulation mechanism that simulates seismic activity to preserve diversity and prevent premature convergence. A single-objective framework is employed to ensure computational efficiency, clarity, and simplicity in model evaluation. Using the CEC 2021 and CEC 2024 benchmark suites, the proposed algorithm is evaluated and shows good accuracy, speed of convergence, and scalability when dealing with single-objective bound-constrained problems. With a total score of 94.668, EDBO places second in CEC 2021, and its efficiency is further supported by its runtime analysis. Statistical evaluations, including standard error analysis and ranking performance, validate its robustness. Additionally, the results of the Friedman and Wilcoxon rank-sum tests demonstrate that EDBO routinely performs better than well-known algorithms such as jellyfish search optimizer (jSOA), modified L-SHADE (mLSHADE-RL), linear success rate-based adaptive differential evolution (L-SRTDE), and reconstructed differential evolution (RDE), with statistically significant improvements (p < 0.05) and achieving a mean Friedman rank of 4.09, EDBO outperform algorithms including ocean water current optimization (OWCO) and many more. These results validate EDBO’s robustness, adaptability, and superior performance across diverse benchmark functions.