Statistical process control charts applied to dairy herd reproduction

The objective of this research was to evaluate the design and performance of statistical process control (SPC) charts when applied to dairy herd reproduction.;A review of the literature involving the application of SPC charts in livestock production revealed that most papers provide an introduction to SPC charts and show examples with livestock performance data. No studies have determined the performance of SPC charts in livestock production.;A stochastic dynamic simulation model of a dairy herd named DASIMO was developed to simulate the performance of SPC charts in dairy herd reproduction. This Monte Carlo model simulates daily performance of individual dairy cows and youngstock over time using functions for milk production, feed intake, disease, reproduction, and animal replacement. DASIMO realistically simulated normal variability in dairy herd reproductive performance.;Various Shewhart and cusum control charts were designed for first service interval, conception interval, estrus detection index, and conception ratio. DASIMO was used to measure the average time to signal on the SPC charts using these four variables when estrus detection efficiency changed. The observed false alarm rate on the SPC charts was occasionally quite different from the target. Cusum charts typically signaled a step change faster than Shewhart charts. The estrus detection index signaled the implemented changes most quickly.;Several topics related to the design of SPC charts for the estrus detection ratio variable were investigated. Equations to calculate the optimal period length for Shewhart X- and P-charts for various changes in estrus detection efficiency were derived. DASIMO was also used to study the effect of varying cusum reference values and false alarm rates on the performance of cusum charts.;The economic value of SPC charts was determined with DASIMO by simulating temporary changes in estrus detection efficiency. Culling policy had a major effect on the economic cost of a temporary change in estrus detection efficiency. Small changes in estrus detection efficiency were shown to be more costly than larger changes because of the increased time to signal on SPC charts.;Stochastic dynamic simulation was demonstrated to be useful for measuring the performance of SPC charts in dairy herd reproduction.