447. Beyond scenarios – resource optimization for dairy cattle breeding programs (MoBPSopti)

Abstract

Modern breeding programs are highly complex with various interdependent parameters and determining the optimal breeding scheme is practically impossible. Therefore, it is a common practice to reduce the optimization problem to a set of scenarios that in turn are deeply analyzed in detail. We are proposing a framework to optimize breeding programs beyond just a comparison of scenarios. We first define the design space of all possible breeding programs and approximated their outcome regarding an objective function using stochastic simulation and local smoothing. Next, the space is reduced to the most promising areas of the design space using various optimization techniques. This process is repeated iteratively until an optimal scheme is obtained. We applied our method to a dairy cattle program with an objective function that includes both genetic gain and genetic diversity conservation under a budget constraint. This study provides more detailed insights into potential approaches of designing and optimizing breeding programs by finding the optimum allocation of resources based on each objective’s overall impact on mid-and longterm breeding goals.