An economic analysis of automatic milking using a simulation model

The economic and logistical consequences of using an automatic milking system (AMS) on a dairy farm in the UK in place of the conventional dairy parlour are examined. A three-phase discrete simulation model is combined with an economic analysis for a range of real farms, herd size of 85-95, and then one of these farms is selected and a sensitivity analysis is conducted. The model is able to simulate milking for cows housed all-year and those that are to be fetched from fields. The model simulates the cows' yield and movement throughout the year, according to their individual lactation curves. The costs and benefits of an AMS for all-year voluntary presentation, for all-year manually fetched (from fields and barn) and for a combination of voluntary presentation from barns in winter, and manually fetched from fields in summer are examined. An analysis of the efficiency of the robot, i.e. time for robot to attach teat cups, time for cow to walk through the system and the time that the robot is available to the cow, is conducted. The utility of the robot is shown to be as important as robot efficiency and the cows will need to be motivated to use the robot for up to 22 h/d. If dairy farmers are to switch from conventional milking to automatic milking, they will need to decide how to deal with the increase in milk yield, by buying more quota or reducing their herd size. The costs of these two strategies are shown. Using current prices and costs, a single-stall robot is compared to a multi-stall robot. The sensitivity of robot cost and lifetime is shown. If the price of quota is low, the robot is competitive for all-year voluntary presentation with the conventional milking parlour, assuming that the robot can be shown to have the same reliability. If automatic milking becomes more popular, the price of the system is likely to fall. Each reduction of pound 1000 in the cost of the one robot and two milking stalls increases the annual relative profit of a farm (used as a baseline cost) by pound 147. (C) 1999 Silsoe Research Institute.