Epsilon Lexicase Selection

Epsilon-lexicase selection is a parent selection method in evolutionary algorithms, particularly genetic programming, aiming to improve the efficiency and effectiveness of finding optimal solutions by prioritizing individuals based on their performance across multiple criteria. Current research focuses on enhancing its performance through techniques like down-sampling and the development of alternative error thresholding methods, such as minimizing variance within error partitions, to address limitations of traditional approaches. These improvements demonstrate significant gains in solving symbolic regression problems, particularly in real-world applications, highlighting the algorithm's practical value and prompting further theoretical investigation into its computational complexity and robustness in handling conflicting objectives.