Increasing economic profit of dairy production utilizing individual real time process data

Automation of feeding and milking enables application of individual cow settings for concentrate allocation and milking frequency. Currently available systems do not derive these settings from real feed efficiencies and milking characteristics of individual cows in their actual situation, but are based on general knowledge. Parameters that characterize true individual cow responses to concentrates and milking intervals can be estimated continuously from real time process data with dynamic linear models. From these parameters individual optimal settings can be determined, such that maximum profit is achieved given the available robot capacity. This approach is developed and succesfully implemented on a research farm for several months. First results show that a substantial gain in milk yield and in profit is possible.