Fuzzy logic applied to tunning mutation size in evolutionary algorithms

Area of Science:

• Computer Science
• Artificial Intelligence
• Optimization

Background:

• Parameter tuning is a critical challenge in Evolutionary Algorithms (EAs).
• Existing EAs often struggle with convergence and getting trapped in local optima due to suboptimal parameter settings.
• Dynamic adaptation of algorithm parameters is essential for robust optimization performance.

Purpose of the Study:

• To propose a novel method for tuning mutation size in Evolutionary Algorithms using Fuzzy Logic.
• To improve the convergence speed and accuracy of EAs towards global optima.
• To enhance the exploration-exploitation balance and prevent premature convergence in EAs.

Main Methods:

• A Fuzzy Logic controller was developed to dynamically adjust mutation size based on historical evolutionary data.
• The Fuzzy Logic controller utilizes past generation data to inform mutation size adjustments.
• The proposed method was tested on benchmark Function Optimization Problems of varying difficulty.

Main Results:

• The Fuzzy Logic-based mutation tuning method demonstrated improved convergence to global optima.
• The approach effectively maintained a balance between exploration and exploitation, reducing the likelihood of local optima entrapment.
• Experimental results on standard benchmark functions confirmed the efficiency of the proposed method.

Conclusions:

• The Fuzzy Logic-based parameter tuning offers an efficient solution for optimizing mutation size in Evolutionary Algorithms.
• This method enhances the robustness and performance of EAs across a variety of optimization problems.
• The approach shows potential for wide applicability in complex optimization tasks.