Precision feeding of gestating sows: considering health, welfare and environmental conditions

Abstract

Better consideration of the variability that exists between gestating sows when it comes to nutrient requirements reduces feed costs without performance deterioration. Nutritional models predicting daily requirements can be coupled with automated systems to deliver an individualized mixture of diets. The variability of nutritional requirements among gestating sows is due to their characteristics (e.g. age, live weight, parity, body condition, stage of gestation), their behaviour (e.g. physical activity, social interactions), their health and the breeding conditions (e.g. ambient temperature, housing system). Sensors can be used to record these factors of variation of nutritional requirements and to help in the detection of occasional disorders at the individual or collective levels (ex. health or technical dysfunction). Thanks to artificial intelligence techniques, which allow the analysis and exportation of heterogeneous data, and the integration of these new behavioural and environmental parameters into nutritional models, it is now possible to meet individual daily requirements more precisely, even in real time.