Modeling of energy ratio index in broiler production units using artificial neural networks

The present study was conducted in Varamin city of Tehran province in Iran. This research addressed the energy analysis of broiler production units. Artificial neural networks (ANNs) were used in order to model energy ratio index on the basis of input energies. Data were gathered from 40 broiler production farms in summer season using a face to face questionnaire approach. The total input and output energy of broiler production were 94783 and 24341.93 MJ (1000 birds)−1, respectively. The energy ratio, energy productivity, specific energy and net energy were 0.26, 0.024 kg MJ−1, 41.16 MJ kg−1 and −70441.07 MJ per thousand birds, respectively. The results revealed that feed, electricity and fuel account for 55%, 29% and 11% of total energy consumption. The developed ANN model with 8-8-13-1 configuration was identified as an optimum topology for predicting the energy ratio. For the best ANN model, the coefficient of determination (R2), mean square error (MSE) and mean absolute error (MAE) were calculated as 0.974, 0.0018 and 0.0337, respectively.