Differential evolution with quasi-reflection-based mutation.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Evolutionary Computation

Background:

• Differential Evolution (DE) is a powerful evolutionary algorithm, but its performance heavily relies on mutation strategies.
• Suboptimal search directions in DE can lead to inefficient exploration and convergence issues.
• Designing effective mutation strategies that consider search direction is crucial for enhancing DE performance.

Purpose of the Study:

• To propose a novel quasi-reflection-based mutation strategy for Differential Evolution.
• To enhance the search direction capability of DE algorithms.
• To improve the overall performance, accuracy, and convergence rate of DE.

Main Methods:

• A quasi-reflection-based mutation strategy was developed and integrated into the DE framework.
• The proposed strategy was combined with the SHADE (Success-History based Adaptive DE) algorithm to create Re-SHADE.
• Performance was evaluated using 30 benchmark functions and the CEC2013 test suite.

Main Results:

• The quasi-reflection-based mutation provides a promising search direction, guiding the DE algorithm more effectively.
• Re-SHADE significantly outperformed existing advanced DE methods in terms of solution accuracy.
• The proposed method demonstrated a superior convergence rate compared to other advanced DE techniques.

Conclusions:

• The quasi-reflection-based mutation strategy is effective in improving DE performance.
• Re-SHADE offers a robust and efficient approach for complex optimization problems.
• The findings confirm the method's effectiveness on standard benchmark test suites.