A Proposal for Applying Bio-inspired Optimization Algorithms: A Comparative Study of Genetic Algorithms and Particle Swarm Optimization for Sensor Selection

Resumo

This paper presents an application proposal and a comparative study between two bio-inspired optimization algorithms: Genetic Algorithms (GA) and Particle Swarm Optimization (PSO). The main objective is to demonstrate the effectiveness of these algorithms in selecting a subset of sensors, aiming to minimize the variance of the collected data. Through the analysis of results from two distinct datasets, this work explores the convergence characteristics, final population distribution, and the profile of the sensors selected by each algorithm. The results indicate that while both algorithms are capable of finding satisfactory solutions, GA tends to achieve better optimization values (lower standard deviation), whereas PSO demonstrates faster convergence. This study contributes to the understanding of the capabilities and limitations of each approach in problems of feature selection and sensing systems optimization.