Digital twins in poultry farming management: A review

Digital twins, defined as dynamic virtual representations of physical systems, offer potential to facilitate real-time monitoring, improve predictive capabilities, and support data-driven decision-making. The application of digital twins in poultry farming management has attracted increasing attention, however, it remains insufficiently explored, particularly in terms of framework development and layer-wise functionalities. This review presents a comprehensive study of digital twin development and components in poultry farming management, addressing the current lack of structured frameworks and systematic analysis in this domain. In this context, poultry farming management is defined as the strategic planning and implementation of a series of farm operations, including monitoring, detection, and prediction, to enhance production efficiency and achieve economic sustainability. Based on the existing literature, a three-layer digital twin framework is proposed, consisting of a data collection layer, a model layer, and an application layer. Within this framework, farm data are first captured and gathered in the data collection layer, then processed and simulated in the model layer, and finally utilized to support decision-making in the application layer. Through a detailed layer-wise analysis, the study highlights the critical role of model calibration and synchronization in ensuring consistency between physical and virtual systems. While offering promising potential, the adoption of digital twin technology in poultry farming faces several challenges, including challenges in real-world deployment, limited system integration, issues related to data quality and ownership, and insufficient solutions to ensure animal health and welfare. Finally, the study summarizes future research directions for digital twin applications in livestock production systems.