On the Improvement of Elite Swimmers Velocity Identification by Using Neural Network Associated to Multiobjective Optimization

Considering that technical skill is the major determinant characteristic of success among competitive swimmers, it is important to coaches to quantify the differences that make one swimmer more efficient than another. One of the most important grants in swimming is the velocity, which can be related to drag forces and provide substantial information about the swimmer technique. The main purpose of this study was to determine the best model to compare swimmers in terms of velocity. In this work, a Radial Basis Function Neural Network (RBF-NN) was used to model the nonlinearity of swim velocity time series. The RBF-NN parameters were adjusted by using four multiobjective optimization methods. The best results in terms of RBF-NN configuration were obtained by the Differential Evolution based algorithms.