An improved FOX optimization algorithm using adaptive exploration and exploitation for global optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• Optimization algorithms face challenges like local minima and balancing exploration/exploitation.
• Existing methods often have numerous hyperparameters, complicating tuning.
• Effective optimization is crucial for diverse real-world applications.

Purpose of the Study:

• Introduce an improved optimization algorithm, Improved FOX (IFOX).
• Enhance the FOX algorithm's performance by refining its core mechanics.
• Address limitations of current optimization techniques.

Main Methods:

• Developed IFOX with a novel adaptive step-size parameter for dynamic exploration/exploitation balance.
• Reduced hyperparameters by removing four parameters (C1, C2, a, Mint).
• Tested IFOX on 20 classical, 61 CEC benchmark functions, and 10 real-world problems.

Main Results:

• IFOX demonstrated a 40% overall performance improvement compared to the original FOX algorithm.
• Achieved superior performance against 16 other optimization algorithms (880 wins, 228 ties, 348 losses).
• Statistical tests confirmed IFOX's competitiveness with state-of-the-art algorithms like LSHADE and NRO.

Conclusions:

• IFOX is a robust and effective optimization algorithm.
• The adaptive step-size mechanism significantly improves performance.
• IFOX shows strong potential for solving complex optimization tasks.