Prospects for Precision Milking Management in Automatic Milking Systems

This paper presents an analysis of the sources of variability in milk yield, milking time and milking interval of cows with voluntary access to a quarter milking Automatic Milking System (AMS). The data used for this analysis included a large data set from a commercial dairy farm using AMS technology. Multiple correspondence analysis and GLMselect were used to explore relationships between variables and select estimation models of the parameters that better identify milking efficiency. This analysis provides some theoretical limits and ranking of precision milking management strategies to improve the capital efficiency of AMS technology. The strategies identified in this analysis were; 1) Increasing the fraction of the day that the AMS box is occupied for milking; 2) Setting a minimum projected milk yield for milking permission; 3) Milking the slowest quarter faster; and 4) Reducing udder prep and attach time. Other substantial value-added propositions for AMS management are to improve the economics of feeding using cow level concentrate feeding and the use of economic models to identify less profitable cows as candidates for dry-off and culling.