Evaluation of a stochastic dynamic replacement and insemination model for dairy cattle

An approach to evaluating results from a stochastic, dynamic model for insemination and replacement of dairy cattle was devised using sensitivity and behavioral analyses. Sensitivity analysis was defined as the quantification of the various outputs resulting from uncertain price and production inputs. Behavioral analysis determined how outputs changed when model specifications were changed. The variation in outputs that were identified by sensitivity analysis was used as an objective measure to assess variation caused by behavioral analysis scenarios. A contemporary model was modified, and a base run was defined using Florida values for input variables. The model specifications that were varied were decision horizon, number of milk production levels, and number of days open classes. Effects of including seasonality of milk production, milk price, and conception rate were examined. The necessary decision horizon was shorter than other works would suggest. Optimal policies were influenced greatly by the number of days open classes, but not by the number of milk production levels. Removal of seasonality of milk production and conception rate resulted in meaningful changes in seasonal patterns of all outputs measured. Our results suggest that model specifications could affect results and should be evaluated objectively when models are being developed.