Optimal Performance and Application for Seagull Optimization Algorithm Using a Hybrid Strategy.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Signal Processing

Background:

• The standard seagull optimization algorithm (SOA) suffers from low search capability and a tendency to get trapped in local optima.
• Enhancing population diversity and balancing exploration-development phases are crucial for robust optimization.

Purpose of the Study:

• To introduce a novel hybrid algorithm, SPSOA, that overcomes the limitations of the standard SOA.
• To improve the global search capability and convergence speed of optimization algorithms.

Main Methods:

• Initialization using Sobol sequences to increase population diversity and ergodicity.
• Integration of a sigmoid function-inspired parameter to balance exploration and development.
• Incorporation of particle swarm optimization (PSO) learning strategy to escape local optima.

Main Results:

• SPSOA demonstrated superior stability, convergence accuracy, and speed compared to other algorithms on 12 benchmark functions.
• The algorithm successfully performed blind source separation on noisy mixed images, outperforming existing methods.
• Experimental results validate SPSOA's effectiveness in both theoretical optimization and practical engineering applications.

Conclusions:

• SPSOA effectively addresses the limitations of the standard SOA, offering enhanced performance.
• The hybrid approach provides a robust solution for complex optimization problems and signal processing tasks like blind source separation.