An improved Grey Wolf Optimizer based on mutation operator, evolutionary population dynamics, and nonlinear...

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristic Computing

Background:

• Standard Grey Wolf Optimizer (GWO) suffers from slow convergence, premature convergence, and an exploration-exploitation imbalance.
• These limitations hinder GWO's effectiveness in diverse engineering applications and real-world optimization tasks.
• Addressing these issues is crucial for improving GWO's efficiency and applicability.

Purpose of the Study:

• To propose a novel variant of the Grey Wolf Optimizer (GWO) named MENGWO.
• To enhance GWO's exploration and exploitation balance and improve convergence speed.
• To validate MENGWO's effectiveness on benchmark functions and engineering design problems.

Main Methods:

• Introduced a mutation operator inspired by Differential Evolution (DE) with adaptive exploration/exploitation switching.
• Incorporated an enhanced Evolutionary Population Dynamics (EPD) mechanism for repositioning underperforming agents.
• Implemented a Nonlinear Population Size Reduction (NPSR) strategy to boost computational efficiency.
• All components feature dynamically adjusted mechanisms based on iteration progression.

Main Results:

• MENGWO demonstrated superior performance compared to standard GWO, Particle Swarm Optimization (PSO), and other GWO variants on CEC2005 and CEC2022 benchmark functions.
• The algorithm showed significant improvements on unimodal, multimodal, and fixed-dimensional multimodal functions across low and high dimensions.
• MENGWO achieved the best performance on 5 out of 7 engineering design problems, indicating strong practical applicability.

Conclusions:

• The proposed MENGWO algorithm effectively balances exploration and exploitation capabilities.
• MENGWO significantly enhances optimization performance, addressing key limitations of the standard GWO.
• The synergistic integration of mutation, EPD, and NPSR strategies makes MENGWO a promising tool for complex engineering applications.