Enhanced crow search algorithm with multi-stage search integration for global optimization problems.

Area of Science:

• Computational Intelligence
• Swarm Intelligence
• Optimization Algorithms

Background:

• The Crow Search Algorithm (CSA) is a swarm intelligence method simulating crow foraging behaviors.
• Standard CSA faces challenges with complex, high-dimensional problems, exhibiting stagnation, slow convergence, and low robustness.
• These limitations stem from CSA's single-stage search relying on random individual interactions.

Purpose of the Study:

• To introduce a Multi-Stage Search Integrated Crow Search Algorithm (MSCSA) to address CSA's deficiencies.
• To enhance the performance of CSA for complex and high-dimensional global optimization problems.
• To improve convergence speed, accuracy, and robustness compared to the original CSA.

Main Methods:

• Introduced chaos and opposition-based learning to enhance initial population quality and ergodicity.
• Implemented a free foraging stage using normal distribution and Lévy flight for local search and accuracy.
• Developed a mixed guiding individual following stage for global search and expanded exploration.
• Incorporated a large-scale migration stage utilizing the best individual to boost diversity and escape local optima.

Main Results:

• MSCSA demonstrated a superior balance between global exploration and local exploitation through its multi-stage approach.
• Experiments on CEC 2017 and CEC 2010 benchmark functions showed MSCSA's competitiveness.
• MSCSA significantly outperformed the original CSA, its variants, and other state-of-the-art meta-heuristics.

Conclusions:

• The proposed MSCSA effectively overcomes the limitations of the standard CSA for complex optimization tasks.
• MSCSA offers enhanced robustness, accuracy, and convergence speed.
• The multi-stage integration strategy proves effective for large-scale, complicated optimization problems.