Exemplar Learning and Memory Retrieval-Based Particle Swarm Optimization Algorithm with Engineering Applications.

Area of Science:

• Computational Intelligence
• Swarm Intelligence
• Optimization Algorithms

Background:

• Particle Swarm Optimization (PSO) is a bio-inspired algorithm known for simplicity and efficiency.
• However, standard PSO struggles with premature convergence and balancing exploration/exploitation due to limited learning and rigid updates.
• These limitations hinder its effectiveness in complex optimization tasks.

Purpose of the Study:

• To propose an enhanced PSO framework, Exemplar Learning and Memory Retrieval-Based Particle Swarm Optimization (EMPSO).
• To address PSO's limitations by integrating novel learning, memory, and adaptation strategies.
• To improve swarm intelligence and overall optimization performance.

Main Methods:

• Developed EMPSO inspired by biological collective behavior.
• Integrated elite exemplar learning for a reliable guidance vector.
• Implemented a memory recall strategy with recency bias for knowledge inheritance.
• Introduced an adaptive position update scheme for dynamic role differentiation.

Main Results:

• EMPSO demonstrated superior performance against six representative algorithms on CEC2017 and CEC2022 benchmark suites.
• The enhanced algorithm showed consistent outperformance across diverse test functions.
• Verified EMPSO's robustness and practical effectiveness through engineering design problems and optimal PMU placement.

Conclusions:

• EMPSO effectively overcomes the premature convergence and exploration-exploitation balance issues of standard PSO.
• The integrated strategies enhance swarm intelligence, leading to improved optimization capabilities.
• EMPSO offers a robust and effective solution for complex optimization challenges in engineering and beyond.