Evaluating a precision feeding decision support system for improving growth performance of growing-finishing pigs on a commercial farm

Abstract

Optimisation of slaughter weight is crucial for efficient farm management in all-in-all-out systems, but growth variability within pig batches complicates uniform marketing. This study aimed to reduce heterogeneity by developing a decision support system (DSS) for precision feeding, improving BW performance, thereby reducing batch variability. A 103-day commercial trial involving 365 pigs compared conventional 3-phase feeding with individual precision feeding (IPF). Two control groups, Control A and Control B (n = 81 each, six pens/group), received diets with stepwise standardised ileal digestible lysine (SID Lys) concentrations (8.80, 9.80, 10.60 g/kg and 9.00, 10.00, 10.80 g/kg, respectively) from traditional feeders, with feed intake recorded manually. In contrast, the IPF group (n = 203, nine pens) utilised robotic feeders to provide individually tailored diets. These were formulated in real-time by blending high (11.83 g/kg) and low (6.59 g/kg) SID Lys feeds. A DSS, integrating a nutritional model, stakeholder directives (minimum and limited daily decreases in the SID Lys concentration), and a qualitative model, calculated each pig’s requirements based on automatically collected real-time BW and feed intake data. Performance metrics were similar across all groups. However, the IPF group (18.55 g/kg) was more efficient in utilising SID Lys, requiring less per kg of live weight gain than Control A (19.67 g/kg) and Control B (19.71 g/kg). When pigs were classified by initial BW – heavy (HBW, IPF: 26; Control A: 23; Control B: 20 animals), moderate (MBW, IPF: 98; Control A: 39; Control B: 41 animals) and light-body-weight (LBW, IPF: 79; Control A: 19; Control B: 20 animals) – the IPF group showed an improvement of 4.2–6.8 kg in growth performance for HBW, and 2.6–4.3 kg in LBW, compared to controls, although not statistically significant. While overall batch variability remained similar (CV: 11.6% IPF, 11.9% Control A, 12.2% Control B), the IPF group was more homogeneous among LBW pigs (9.5%) compared to controls (11.5% and 13.8%). Greater HBW variation in IPF group balanced overall variability. Although direct feed cost savings and nitrogen excretion reductions were not achieved – attributed to technical feed distribution issues in the final phase and higher CP baselines in the experimental diets – an economic estimation revealed that the system’s profitability was driven by output maximisation. In conclusion, the DSS proved feasible for real-time commercial application, successfully enhancing nutrient utilisation efficiency and optimising the growth of animals at the extremes of the population distribution.