A self-learning particle swarm optimizer for global optimization problems.

Area of Science:

• Computational intelligence
• Optimization algorithms
• Swarm intelligence

Background:

• Particle Swarm Optimization (PSO) is effective for global optimization.
• Existing PSO algorithms often use a single, uniform learning pattern for all particles.
• This uniformity can limit a particle's adaptability to complex search spaces.

Purpose of the Study:

• To introduce a novel algorithm, the Self-Learning Particle Swarm Optimizer (SLPSO).
• To enhance the adaptability and intelligence of individual particles within a swarm.
• To improve the performance of PSO in solving global optimization problems.

Main Methods:

• Developed SLPSO with each particle possessing four distinct strategies.
• Implemented an adaptive learning framework for individual-level strategy selection.
• Evaluated SLPSO on 45 diverse test functions and two real-world optimization problems.

Main Results:

• SLPSO demonstrated superior performance compared to several peer optimization algorithms.
• The adaptive learning framework enabled particles to select optimal strategies based on local fitness landscapes.
• The novel approach improved the ability to handle complex optimization scenarios.

Conclusions:

• SLPSO offers a significant advancement over traditional PSO methods.
• Individualized adaptive learning enhances particle swarm optimization effectiveness.
• The proposed SLPSO is a promising tool for tackling complex global optimization challenges.