Dynamic programming applied to the dairy cow replacement decision

This research addresses the question ‘When should dairy cows be replaced?’ A stochastic dynamic programming model was used to derive optimal replacement policies based on age, time of calving, and milk production performance. The policy indicated whether an animal in a particular state should be kept or culled. A method of calculating the correlation coefficient between the average of κ and κ+1 performances was devised, which enabled the bivariate normal distribution to be used in determining the probability of transitting from one performance class to another. The results of the study suggest that the main emphasis on culling should be on the first lactation performance.