Análise de confiabilidade aplicada na otimização de sistemas de produção animal

Abstract

Although Reliability Engineering focuses on the study and discussion of the life of inanimate objects, such as equipment, components, items, sub-items or systems, these same concepts can be applied in other areas, such as survival analysis in the optimization of animal production - object of study of this work in which the target variable, resulting from a 2 x 3 factorial experiment, describes the number of days elapsed between birth and the first posture, event of interest, from yellow, blue and red quails and submitted to two levels of diet. It was attempted to select the best probabilistic model that describes in a reasonable way the behavior of the age, in days, until the first posture, as well as to estimate punctually and by interval, the average time of the first posture and, finally, to evaluate the effect of the treatment diet in the bloodlines. The analytical and graphical procedures, used to evaluate the best fit, show that the Generalized Gama distribution in most groups is the most adequate to describe the target variable of interest. With a probabilistic model capable of describing effectively the phenomenon of interest, there are significant implications in the decision process regarding animal management and consequently in the optimization of the production system. In this modeling, the yellow bloodline presented a shorter posture time around 40 days and the diet 1, independent of the lineage, provided a shorter posture time for all treatments.