Multi-strategy synthetized equilibrium optimizer and application.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Machine Learning

Background:

• Equilibrium Optimizer (EO) faces challenges with search ability, computational complexity, and operability in complex optimization.
• Existing algorithms require improvements for enhanced performance in demanding computational tasks.

Purpose of the Study:

• To propose an improved Equilibrium Optimizer (EO) algorithm, termed Multi-Strategy on Updating Synthetized EO (MS-EO).
• To enhance the search ability, reduce computational complexity, and improve the operability of the EO algorithm.

Main Methods:

• Introduced a simplified updating strategy to enhance operability and reduce complexity.
• Integrated an information sharing strategy with dynamic tuning for improved exploration.
• Incorporated migration and golden section strategies for enhanced search capabilities.
• Implemented an elite learning strategy to strengthen exploitation ability, balancing exploration and exploitation.

Main Results:

• MS-EO significantly outperformed the original EO and other state-of-the-art algorithms on CEC2013 and CEC2017 complex function test sets.
• Demonstrated superior search ability, running speed, and operability compared to existing methods.
• Showcased advantages in feature selection tasks across multiple datasets.

Conclusions:

• The proposed MS-EO effectively balances exploration and exploitation, offering significant improvements over the standard EO.
• MS-EO presents a robust and efficient alternative for complex optimization problems and feature selection applications.