Boosting the performance of Differential Evolution Metaheuristic.

Abstract

Optimization is a critical tool in the scientific community, driving advancements and solving complex problems across various domains. Among the various optimization techniques, metaheuristics have gained prominence for their flexibility and effectiveness. Differential Evolution (DE), a leading metaheuristic algorithm, is widely used due to its efficiency and robustness in handling diverse optimization challenges. DE has been uccessfully applied in numerous real-world applications, including engineering design, machine learning, and financial modeling. Despite its strengths, DE faces many challenges such as balancing exploration and exploitation and maintaining performance in highdimensional spaces. The performance of DE is largely attributed to the arameters and operators used. Motivated by this, we combine ideas from different existing approaches to enhance DE’s capabilities. This includes modifying the basic fixed population size and employing various operator strategies with parameter control techniques. By doing this, we aim to improve the efficiency of DE across diverse optimization scenarios. Through extensive experimentation using the BBOB benchmarking environment, the new approach demonstrates promising results.