Optimal replacement policies for dairy cows based on daily yield measurements

Area of Science:

• Dairy science
• Animal husbandry
• Operations research

Background:

• Traditional Markov decision process (MDP) models for dairy herds use monthly stages, limiting daily decision-making.
• Modern dairy management systems provide daily milk yield data, creating a need for more granular decision support.

Purpose of the Study:

• To develop a hierarchical Markov decision process (MDP) model with daily stages for the dairy cow replacement problem.
• To create a model that integrates with modern herd management systems for daily decision support.

Main Methods:

• Formulated a hierarchical MDP with daily stage lengths for dairy cow replacement.
• Employed Bayesian updating to predict individual cow performance.
• Estimated model parameters using herd-specific data, including lactation curves.

Main Results:

• Developed a novel daily MDP model applicable to dairy cow replacement.
• The model leverages daily milk yield measurements for enhanced decision-making.
• Incorporated herd-specific data for accurate performance predictions and economic evaluations.

Conclusions:

• The developed hierarchical MDP model enables daily replacement decisions in dairy herds.
• This approach supports integration into modern herd management systems.
• Utilizing daily data and herd-specific parameters improves the precision of dairy farm management.