Demands and opportunities for operational management support.

It is generally recognised that with automatic milking operational farm management is considerably changed compared to conventional machine milking, and is a key issue to fulfil requirements in practise. Automatic milking causes a reduction of physical labour, but an increase in decision making tasks of the dairy farmer. Decision making is supported by information obtained with sensors. Each manufacturer has different solutions for management support and performance of automatic tasks. The concept of management by exception and the distinction between normal and abnormal parameter values are discussed. Difficulties that arise when detection of abnormalities is based on correlated traits are outlined. Characteristics of sensors and their consequences are discussed for both automated decision making and decision support. The paper also describes differences between conventional and automatic milking with respect to operational farm management. A clear difference with conventional milking is that milking intervals have to be controlled for individual cows. Feeding strategy is a key element in this, especially when grazing is applied. With regard to health, in general much more information is available, especially detection of mastitis partly depends on abnormalities detected with sensors. However, visual inspection of the animals remains an important method to control health. Because milkings are unattended, regularly abnormalities have to be checked. Due to the limited reliability of the alarms, and the fact that these are not yet integrated, this requires specific skills from the farmer. Furthermore the automatic milking system (AM system) has to be maintained and its cleaning and functioning must be controlled more or less constantly. In order to be able to comply with legislation some of the currently existing regulation needs to be adapted, avoiding double standards. Farmers must have affinity with automation. They have to work with secure procedures, for instance first enter treatment data in the computer and then treat sick cows, and respond as should on alarms for system and animals. Udders and teats must be kept clean. Farmers using AM systems in general are satisfied with the current possibilities. Despite this, not all of the demands for automatic milking are yet fulfilled by the current systems but this is also due to inadequate legislation. Especially automatic separation of abnormal milk and secured teat cleaning should be realised. Furthermore, monitoring of the animals and equipment can be further improved. With these improvements automatic milking certainly has advantages with respect to milk quality and food safety compared to conventional milking.