Modelagem fuzzy do desempenho de frangos de corte, criados de 1 a 21 dias, submetidos a estresse térmico

Abstract

Faced with the complexity with which the parameters of broiler production environment relate, artificial intelligence may be used to aid in the decision making of broiler farming, seeking an increase in productivity and product quality. Thus, the objective of this study was to verify the performance of 1-21 days broilers submitted to thermal stress in the second week of life, with varying intensities and durations in conditioned wind tunnels, and using the results to validate the fuzzy logic model. The performance was analyzed at the Animal Environment Laboratory of the Universidade Federal de Lavras, in which all thermal variables were controlled, varying only the dry bulb temperature during the second experimental week. Thus, it was possible to evaluate feed intake (g), weight gain (g) and feed conversion (g g-1) of the broilers when subjected to different intensities and durations of thermal stress by low and high temperatures. A mathematical model was developed using the fuzzy set theory to predict the performance of broilers regarding stress temperature (°C) and duration of the stress (days). The validation of the proposed model was performed comparing the experimental data with the results obtained by the modeling. Therefore, with the results obtained, the fuzzy modeling may be used to predict the performance of broilers when subjected to thermal stress in the second week of life.