Improved moth-flame algorithm based on cat chaotic and dynamic cosine factor.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Machine Learning

Background:

• The standard Moth-Flame Optimization (MFO) algorithm faces challenges with insufficient population diversity and unbalanced search capabilities in complex optimization tasks.
• Addressing these limitations is crucial for improving the efficiency and effectiveness of optimization algorithms.

Purpose of the Study:

• To propose an Improved Moth-Flame Optimization (IMFO) algorithm to overcome the shortcomings of the traditional MFO.
• To enhance the performance of MFO in solving complex function optimization problems and engineering applications.

Main Methods:

• Introduced cat chaotic mapping for initial population diversity.
• Incorporated cosine inertia weight to balance global and local search.
• Integrated particle swarm optimization memory for faster convergence.
• Applied Gaussian mutation strategy to prevent local optima.

Main Results:

• IMFO demonstrated significantly improved population diversity and balanced search capabilities.
• Simulation experiments on 11 benchmark functions showed higher accuracy and stability compared to other MFO variants and classical algorithms.
• Validation through Kernel Extreme Learning Machine (KELM) optimization in an engineering example confirmed superior performance.

Conclusions:

• The proposed IMFO algorithm effectively addresses the limitations of the standard MFO.
• IMFO offers enhanced accuracy, stability, and convergence speed for optimization problems.
• The algorithm shows promising potential for practical engineering applications, particularly in machine learning model optimization.