EVOLUTIONARY ALGORITHMS FOR SOLVING COMPLEX OPTIMIZATION PROBLEMS IN DATA SCIENCE

Abstract

This study presents a hybrid multi-evolutionary algorithm that integrates the strengths of Genetic
Algorithms (GAs) and Evolutionary Strategies (ESs) to effectively tackle complex optimization problems in data
science. The proposed GA-ES approach leverages the exploratory capabilities of GAs alongside the rapid
convergence attributes of ESs, facilitating a balanced navigation of diverse solution landscapes. Experimental
results demonstrate that this hybrid methodology outperforms traditional optimization techniques, particularly
in scenarios characterized by multiple local optima. Additionally, the interchange of top-performing individuals
between the two algorithms enhances optimization efficiency and leads to superior solutions within a shorter
computational timeframe. The findings highlight the potential of evolutionary algorithms as robust tools for
addressing intricate optimization challenges. Future research is encouraged to refine the hybrid framework and
investigate its application across a wider array of real-world data science problems, particularly those
necessitating adaptive optimization strategies, thereby contributing to advancements in data-driven decisionmaking
processes.