Modelos de Regressão Simbólica Através de Biased Random-Key Genetic Programming em Aplicações na Física

Abstract

This work focuses on the application of the meta-heuristic Biased Random-Key Genetic Programming (BRKGP) in the context of the Symbolic Regression Problem. The goal is to approximate the algebraic form of an unknown equation based on numerical inputs and outputs. The BRKGP algorithm, based on genetic programming, plays a crucial role in this process by optimizing the generation of symbolic expressions. The study presents preliminary results and compares accuracy with different sample sizes derived from physics problems. The research aims to explore the application of BRKGP in symbolic regression, seeking to contribute to the validation of physical equations and promote advancements. The importance of improvements in computational techniques is emphasized, highlighting challenges such as indeterminacies and extreme numbers. Notable results indicate comprehensive applications in various fields. The study concludes by emphasizing the efficiency of the innovative approach and pointing out challenges to be addressed in future studies, including the refinement of data treatment strategies and the elimination of indeterminacies.