Underestimation-Assisted Global-Local Cooperative Differential Evolution and the Application to Protein Structure...

Area of Science:

• Computational Biology
• Optimization Algorithms
• Bioinformatics

Background:

• Differential evolution (DE) algorithms utilize various mutation strategies for optimization.
• Cooperation between multiple DE strategies can potentially enhance performance.
• Existing DE methods may lack simultaneous effectiveness and efficiency in complex problems.

Purpose of the Study:

• To propose an underestimation-assisted global and local cooperative DE algorithm.
• To enhance both the effectiveness and efficiency of differential evolution.
• To improve performance in complex optimization tasks like protein structure prediction.

Main Methods:

• Implementing a two-phase approach (global exploration and local exploitation) within each generation.
• Utilizing multiple DE strategies for generating trial vectors in the global phase.
• Employing an adaptive underestimation model with self-adapted slope control for trial vector evaluation.
• Designing better-based strategies for the local exploitation phase.
• Incorporating adaptive parameter control for DE.

Main Results:

• The proposed cooperative DE algorithm demonstrates superior performance compared to competitors.
• Experimental results on benchmark functions and CEC test sets validate the approach.
• The algorithm shows significant effectiveness and efficiency in protein 3D structure prediction.

Conclusions:

• The underestimation-assisted global and local cooperative DE effectively balances exploration and exploitation.
• The cooperative strategy enhances the overall performance of differential evolution.
• The proposed method offers a robust solution for complex optimization problems, particularly in bioinformatics.