Abstract: Traditionally, cereal grains are the major source for the non-ruminant livestock feeds. Unfortunately in most developed countries there is always a shortfall between the production and demand for cereal grains. There is an intense competition between humans, livestock and industrial users for the little quantity of grains that are available. In the production of poultry birds feed composition is an important factor and the composition determines the consumption and conversion after other proper conditions have been met. In this study artificial neural network techniques is used to predict the weight of poultry birds (broilers) reared with different feed diets where maize has been supplemented with sorghum and an exogenous enzyme (â-glucanase). Various neural network models were tested. The best neural network model developed was a 4-input, 1-output Generalized Feed Forward Neural Network (GFFNN) having 1-hidden layer with 3 processing elements trained using the Back propagation of errors method. The results obtained confirmed that feed compositions comprising of mainly maize was still the best for poultry production, however the neural network was able to predict the optimal bird weights derivable from poultry diets where the maize content has been supplemented with sorghum dust.
A.S. Ajakaiye , A.B. Adeyemo , A.O. Osofisan and O.P.A. Olowu , 2006. Analysis of Poultry Birds Production Performance using Artificial Neural Networks. Asian Journal of Information Technology, 5: 522-527.