Gaussian Perturbation Specular Reflection Learning and Golden-Sine-Mechanism-Based Elephant Herding Optimization for...

Area of Science:

• Computational Intelligence
• Metaheuristic Optimization
• Algorithm Improvement

Background:

• Elephant Herding Optimization (EHO) is recognized for its simplicity but struggles with slow convergence and low accuracy.
• Existing metaheuristic algorithms often face challenges in balancing exploration and exploitation for complex problems.

Purpose of the Study:

• To propose an enhanced Elephant Herding Optimization algorithm, SRGS-EHO, addressing limitations in convergence speed and solution accuracy.
• To improve the global exploration and local exploitation capabilities of the EHO algorithm.

Main Methods:

• Introduced specular reflection learning and Gaussian perturbation to enhance initial population diversity and convergence.
• Integrated a golden sine mechanism to optimize patriarch position updates for better global optimum seeking.
• Evaluated SRGS-EHO on 23 benchmark functions and two engineering design problems.

Main Results:

• SRGS-EHO demonstrated significant improvements in convergence speed and solution accuracy compared to standard EHO.
• Statistical tests (Wilcoxon, Friedman) confirmed SRGS-EHO's superiority over eight other metaheuristic algorithms.
• The algorithm successfully solved constrained engineering problems, indicating practical applicability.

Conclusions:

• The proposed SRGS-EHO algorithm effectively overcomes the limitations of traditional EHO.
• SRGS-EHO offers enhanced global exploration and local exploitation, leading to better optimization performance.
• The algorithm shows promise for solving real-world engineering optimization challenges.