A multi-sample particle swarm optimization algorithm based on electric field force.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Swarm Intelligence

Background:

• Particle swarm optimization (PSO) algorithms often suffer from premature convergence, limiting their effectiveness.
• Existing PSO variants struggle to maintain diversity and adapt to complex search spaces.
• The need for robust optimization techniques that balance exploration and exploitation is critical.

Purpose of the Study:

• To address the premature convergence issue in particle swarm optimization.
• To propose a novel multi-sample particle swarm optimization (MSPSO) algorithm incorporating electric field dynamics.
• To enhance the search efficiency, convergence accuracy, and practical applicability of PSO.

Main Methods:

• Introduced electric field force to induce diverse particle behaviors within the swarm.
• Developed an electric field force-based comprehensive learning strategy (EFCLS) using attractive and repulsive samples.
• Implemented a segment-based weighted learning strategy (SWLS) for global learning sample construction.
• Employed adaptive parameter adjustment based on population status.

Main Results:

• MSPSO demonstrated superior performance across sixteen benchmark functions and eight PSO variants.
• The algorithm achieved higher accuracy, particularly in high-dimensional optimization problems.
• MSPSO exhibited a faster convergence rate compared to existing methods.
• Validation on a real-world problem confirmed the practical utility of MSPSO.

Conclusions:

• The proposed MSPSO algorithm effectively mitigates premature convergence in particle swarm optimization.
• The integration of electric field force and novel learning strategies significantly enhances search efficiency and accuracy.
• MSPSO offers a promising and practical solution for complex optimization tasks, including those in high-dimensional and real-world scenarios.