Clean Genetic Algorithm Architecture for Improved Modularity and Extensibility
Zoran Janković, Boban Vesin · 2023
This paper delves into the potential of incorporating SOLID and clean architecture principles into the development of evolutionary algorithms, specifically focusing on a genetic algorithm. The introduced approach presents an innovative design and implementation, offering notable enhancements in extensibility and modularity, thereby contributing to improved testability and maintainability within the domain of evolutionary algorithms. The study demonstrates how the proposed design enables effortless extensions of operations within the genetic algorithm, such as parent selection, crossover, and mutation. These enhancements result from strict adherence to the SOLID and clean architecture principles. To demonstrate the practicality of this approach, the principles have been applied to a Python library called "GAdapt." This paper highlights the constructive influence of introduced concepts on the development of genetic algorithm and the improved flexibility and ease of maintenance inherent in this approach.