A synchronous-asynchronous particle swarm optimisation algorithm.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Swarm Intelligence

Background:

• Traditional synchronous particle swarm optimisation (S-PSO) excels at exploitation but can be slow.
• Asynchronous particle swarm optimisation (A-PSO) enhances exploration but may lack convergence efficiency.

Purpose of the Study:

• To develop a hybrid particle swarm optimisation (PSO) algorithm combining synchronous and asynchronous update strategies.
• To leverage the strengths of both S-PSO and A-PSO for improved optimization performance.

Main Methods:

• Introduction of the synchronous-asynchronous PSO (SA-PSO) algorithm, dividing particles into synchronously updated groups.
• Utilizing group best and swarm best to guide the search process.
• Comparative performance analysis against S-PSO and A-PSO on unimodal, multimodal, and real-world optimization problems.

Main Results:

• The proposed SA-PSO algorithm demonstrated consistent and strong performance across various benchmark functions.
• Statistical analysis confirmed the effectiveness of the hybrid approach in balancing exploration and exploitation.
• SA-PSO showed competitive or superior results compared to both S-PSO and A-PSO.

Conclusions:

• The synchronous-asynchronous PSO (SA-PSO) effectively integrates synchronous and asynchronous update mechanisms.
• SA-PSO offers a robust and consistent optimization strategy for diverse problem landscapes.
• The hybrid approach represents a significant advancement in particle swarm optimisation techniques.