Dairy barns, robots and queueing networks

In an experimental farm in Duiven in the Netherlands the agricultural research center IMAGDLO is investigating the behaviour of the cows in an RMB. Based on extensive measurements and observations, it was in [2] concluded that it is necessary to incorporate the stochastic behaviour of the cows in the design of an RMB. Another aspect, which makes this design complex, is the interaction between the facilities in the barn: increasing the capacity of bottleneck facilities will shift queues and alter the location of bottlenecks, possibly forcing the designer to increase the capacity elsewhere. The concept of the closed queueing network (CQN) seems to be very appropriate for modelling and analysing an RMB. It covers both the random behaviour of the cows and the interaction between the queues. It also supports a systematic analysis of the tradeoffs between economic and cow-social factors. The CQN model is widely used in the communication systems and production systems areas. The present application area, the design of RMBs, is new. Based on the CQN the Java applet Cow has been developed, which allows the user to easily and efficiently evaluate and compare alternative designs. We will first describe the milking robot and the dairy barn in more detail. Then we introduce our queueing network model of the barn. Finally we discuss the Java applet and show that this tool is very useful for the design of a barn. The milking robot The milking robot is shown in figure 1. Milking robots are different from the ordinary milking machines in one crucial aspect: the robot uses sensors to find the teats of the cow and then connects the cups to the teats with a robot arm. Milking robots, their operation and costs have been extensively reviewed in e.g. [1, 6, 7]. So milking is done without intervention of the farmer. This saves the farmer serious amount of labour. Further, it makes it possible to go from milking twice a day to three or even more times a day. When cows are milked three times a day their production is increased by about 15 percent. Figure 1: A milking robot.