Abstract
The nutritional requirements of growing-finishing pigs are based on a complex interplay of parameters like feed intake (FI), metabolism, and other environmental conditions. We developed an individual-based model to estimate the growth performance and nutritional requirements of growing-finishing pigs. Our model implementation follows the principles and equations published by the 2012 Swine National Research Council (NRC) model, focusing on reflecting the attributes of each pig and its interaction with the environment. The proposed swine nutrition system (SNS) utilized an agent-based model framework and was developed using NetLogo to simulate the dynamics of pig nutrition systems. The main factors incorporated in the model were body weight (BW) gain (BWG), start and finish BW, and the sex of the pigs. Results showed that the model can accurately estimate foundational parameters of pigs' biological growth, including BW, FI, metabolizable energy intake (MEI), protein deposition (Pd), and lipid deposition (Ld). In addition, the daily requirements, such as amino acid, calcium, and phosphorus requirements, were calculated separately for each pig. The model accurately demonstrates known differences in pig growth characteristics, such as greater Ld in barrows, greater daily protein accretion, and dietary standardized ileal digestible lysine requirements in boars. The proposed model was evaluated by comparing SNS predictions' correlation and coefficient of determination (r2) to those of the 2012 Swine NRC model's predictions. The simulations were conducted for 500 pigs to demonstrate the repeatability of the SNS. The average, standard deviation, and the 95% confidence interval were obtained over the growth period of the pigs for the parameters, such as BWG, BW, Pd, Ld, MEI, and FI. The correlation between BWG from the SNS and BWG from the NRC was r2 = 0.84 for gilts, r2 = 0.93 for barrows, and r2 = 0.93 for boars. Also, the direct comparison between the SNS and the 2012 Swine NRC indicated an r2 > 0.99 for the three sexes. SNS enabled the simulation of individual animal behavior, nutrient partitioning, and variability in growth performance, which are capabilities not afforded by traditional aggregate models. Thus, real-world implementation of SNS might improve feeding and management strategies within commercial swine production systems, leading to greater production efficiency and sustainability in the pork sector.