On the performance improvement of Butterfly Optimization approaches for global optimization and Feature Selection

Area of Science:

• Computational Intelligence
• Metaheuristics Optimization
• Swarm Intelligence

Background:

• The Butterfly Optimization Algorithm (BOA) is a nature-inspired metaheuristic algorithm.
• BOA can be prone to getting trapped in local optima, limiting its exploration and exploitation balance.
• Enhancements are needed to improve BOA's efficiency and effectiveness.

Purpose of the Study:

• To develop improved versions of the Butterfly Optimization Algorithm (BOA).
• To enhance BOA's ability to escape local optima and balance exploration/exploitation.
• To evaluate the performance of enhanced BOA versions on benchmark functions and real-world problems.

Main Methods:

• Developed three enhanced versions of BOA: Opposition-Based Strategy (OBS), Chaotic Local Search (CLS), and a combined CLS-OBS (CLSOBBOA).
• Compared the proposed BOA versions against original BOA and other metaheuristics (GWO, MFO, PSO, SCA, WOA).
• Tested algorithms on CEC 2014 benchmark functions, four engineering design problems, and a feature selection task using UCI datasets.

Main Results:

• The third version, CLSOBBOA, integrating both Opposition-Based Strategy and Chaotic Local Search, showed superior performance.
• CLSOBBOA achieved better results in terms of speed and accuracy compared to other algorithms.
• The enhanced BOA versions demonstrated effectiveness in solving complex optimization and feature selection problems.

Conclusions:

• The proposed CLSOBBOA significantly improves upon the original BOA.
• The integration of Opposition-Based Strategy and Chaotic Local Search is effective for enhancing metaheuristic performance.
• CLSOBBOA offers a promising approach for addressing complex optimization challenges in engineering and data science.