Interests and limits of the representation of individual variability in dynamic herd models

Abstract

The variability of individual responses is a key element to understand and evaluate herd performance. It has a biological basis(responses generated by physiological regulations and genotype) and a decisional basis (different management practices involving differentherd organisation levels). This review analyses if current dynamic herd models were appropriate to study individual variabilityand produce knowledge to benefit from it. After analysing the type of biological responses represented in models and how theseresponses are modelled and then the type of management practices and the herd levels involved to perform these practices, the conclusionpresents the models’ shortcomings to study individual variability. They emerge from a partial view of the biological basis (onlyone type of biological response represented, no response to feeding and no regulations) and of the decisional one (no individual information,a single type of practices and no link between farmer’s project, groups of animals and practices). As a result, simulating theemergence of individual variability from herd operation requires developing agent-based herd model with two essential characteristics:first, representing the animal as a regulated organism in a nutritional environment and second, formalizing groups of animalsas management entities, targeted by a set of feeding, reproductive and replacement practices.