Designing the optimal robotic milking barn, part 1: Quantifying facility usage

The aim of this paper is to explore the stochastic nature of the usage of facilities in a robotic milking barn, independent of the barn layout. It presents experimental data obtained by monitoring 10 dairy cows over a period of 18 days. To minimize restrictions to the access of cows to the facilities, the barn contained less than half the number of the cows for which it was designed. Under these conditions of maximum availability of facilities, the intensity and sequence of facilities usage were studied. The access to all the facilities in the barn can be approximated by an exponential distribution where the values for the exponential rate θ for access to concentrates, forage, water, cubicles and milking, respectively, are: θ_conc. of 13.15, θ_forage of 8.77, θ_water of 12.06, θ_cubicles of 8.08, and θ_milking of 15.11. The 'flow' of the cows between the facilities was expressed in a transition matrix. The first priority of cows crossing from the resting area to the feeding area through the milking robot stalls was concentrate feeding (91.4% of the events). The occupation rate (cows/positions) of forage lane or cubicles was less than the milking-parlour situation. Robotic milking evened out the usage of all the facilities in the barn throughout day and night to a continuous-time stochastic process. Quantifying this stochastic process under these conditions of maximum availability of facilities opens up the possibility to allocate facilities optimally, based on known facility usage by the animals.