An enhanced seasons optimization algorithm for numerical optimization and engineering design.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• Standard Seasons Optimization (SO) shows promise but struggles with exploration-exploitation balance, leading to premature convergence.
• Existing optimization algorithms often face challenges in achieving high-quality solutions and efficient convergence.
• Nature-inspired algorithms are increasingly vital for complex numerical and engineering optimization problems.

Purpose of the Study:

• To introduce the Enhanced Seasons Optimization (ESO) algorithm, addressing the limitations of the standard SO.
• To improve search performance, solution quality, and convergence speed in optimization tasks.
• To evaluate ESO's effectiveness against a wide range of established and novel optimization algorithms.

Main Methods:

• ESO integrates four novel operators: root spreading for local exploitation, wildfire for population diversity, refined competition/resistance for quality and convergence, and opposition-based learning to avoid local optima.
• Comparative analysis involved 25 numerical optimization functions and 4 engineering design problems.
• Performance was benchmarked against foundational (PSO, DE), top-performing (CMAES, LSHES, RRTO, ALA, THRO), and other nature-inspired algorithms (SO, HO, CBKA).

Main Results:

• ESO significantly outperformed the standard SO algorithm across tested benchmarks.
• ESO demonstrated competitive or superior performance compared to all counterpart optimizers in solution quality and convergence.
• ESO ranked first in 16/25 numerical functions and 3/4 engineering problems, achieving the best average rank in the Friedman test for numerical benchmarks.

Conclusions:

• The Enhanced Seasons Optimization (ESO) algorithm effectively overcomes the limitations of the standard SO, offering improved search performance and solution quality.
• ESO exhibits robust performance, particularly on shifted, composite, and high-dimensional (1000-D) numerical problems, and in scalability analysis.
• ESO represents a significant advancement in metaheuristic optimization, providing a powerful tool for complex computational problems.