Uma abordagem interdisciplinar para obtenção da melhor função aproximada para ajuste de dados

Abstract

In this work, important concepts from Exact Sciences, Computing and Engineering act in an interdisciplinary way to obtain the best approximate solution for a physical experiment. The aim of this paper is to show how different theories can be used to solve problems in engineering. More specifically, the Least Squares method (studied in Linear Algebra), the uncertainties and interval operations (studied in Interval Mathematics) and the Python language (studied in computation) will be used to solve an interval system that provides the best approximate function that adjusts a data set from a physical experiment in which a car was moving at zero acceleration under a horizontal air track. By doing this type of approach using Interval Mathematics, we seek to infer how the uncertainties arising from the experiment, as well as the errors generated by the representations and operations of computer numbers, interfere with the obtained result. For this, it was necessary to use the Python for Extended Scientific Computing (Python-XSC) library, which is based on the interval arithmetic structure and provides functions for the resolution of interval linear systems. The application of this study proved to be very efficient and easy to use, besides encouraging engineering students to seek innovative solutions through the interdisciplinarity of the theories studied in the area.