Modelagem fuzzy como ferramenta para predição do ganho de peso diário para frangos de corte

Abstract

Many varied empiric mathematical models have already been developed to predict broiler chicken growth. However, newly developed tools, such as fuzzy modeling, may facilitate resolutions of problems. The objective of this research is to develop a computational mathematical model based on fuzzy logic to predict the daily weight gain (GPD) of broiler chickens, and to compare it to statistical models from the literature. Our results showed that a standard deviation of 0.9 g.day-1 was estimated using fuzzy model compared to 2.02 g.day-1 from the empirical model. However, the standard deviation was 8.6 g.day-1 when field data was used in the fuzzy model, and 9.8 g.day-1 for the empirical model. The proposed fuzzy modeling showed better precision compared to the empiric model. However, the results were not as good when experimental field data was used (R² = 0.5677).