Markov decision processes in the optimisation of culling decisions for Irish dairy herds
Haran, Peter Noel
(1997)
Markov decision processes in the optimisation of culling decisions for Irish dairy herds.
Master of Science thesis, Dublin City University.
This thesis presents the development of a decision support system to optimise replacement and insemination decisions in Irish dairy herds under market conditions. The technique used was developed specifically for animal replacement problems, and is known as a ‘Hierarchic Markov Process’. The model optimises culling decisions on the basis of lactation, production level, calving date, fertility and calving interval.
Production, involuntary culling rates, feed costs, carcass values and other economic factors are allowed to vary according to the traits of a particular animal. The output from the model is a series of retention payoffs (RPO) upon which culling decisions are based. RPO is the expected future return in keeping a cow for an additional stage rather than replacing her with a heifer.
A study of culling rates in commercial dairy herds was also carried out and the effect of culling strategies on the genetic level of dairy herds was investigated using Monte Carlo simulation.