A New Hybrid MGBPSO-GSA Variant for Improving Function Optimization Solution in Search Space

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Nature-Inspired Computing

Background:

• Traditional optimization algorithms often struggle with balancing local exploitation and global exploration.
• Particle Swarm Optimization (PSO) excels in exploitation, while Gravitational Search Algorithm (GSA) is strong in exploration.
• Combining these algorithms can potentially create a more robust optimization framework.

Purpose of the Study:

• To develop a novel hybrid nature-inspired optimization approach by integrating Mean Gbest Particle Swarm Optimization (MGBPSO) and Gravitational Search Algorithm (GSA).
• To achieve an automatic balance between local and global search capabilities.
• To evaluate the performance of the proposed MGBPSO-GSA approach on benchmark functions and real-world problems.

Main Methods:

• Developed a hybrid algorithm, MGBPSO-GSA, by combining the exploitation strengths of MGBPSO with the exploration capabilities of GSA.
• Tested the algorithm's performance on unimodal, multimodal, and fixed-dimension multimodal benchmark functions.
• Validated the approach by applying it to the Iris dataset, Heart dataset, and economic dispatch problems, comparing it against existing metaheuristics.

Main Results:

• The MGBPSO-GSA approach demonstrated a significant capability for automatic balance between local and global search.
• Empirical results showed that the hybrid approach significantly outperformed several other metaheuristics.
• Superior performance was observed in terms of solution stability, solution quality, local and global optimum finding, and convergence speed.

Conclusions:

• The MGBPSO-GSA hybrid algorithm offers a powerful and effective new tool for complex optimization tasks.
• The integration successfully synthesized the strengths of MGBPSO and GSA, leading to enhanced performance.
• The approach shows promise for applications in machine learning (datasets) and engineering (economic dispatch).