Optimal slaughter pig marketing with emphasis on information from on-line live weight assessment

One of the most labor intensive tasks of a traditional slaughter pig production is weighing of pigs for marketing, and in several recent projects methods for automatic assessment of live weights are developed. For an optimal utilization of the resulting weighing data, a decision support system for optimization of the marketing policy is needed. In this paper, a pen level model intended as the core of such a decision support system for optimization of slaughter pig marketing is presented. The model is based on a hierarchical Markov process and emphasis is put on definition of the state space in such a way that the observations from the online live weight estimation may serve as input to the decision support system and thus improve the precision of the underlying predictions of future growth through a learning algorithm based on Kalman filtering in a dynamic linear model. The aim is to become able to inform the farmer on the number of pigs being ready for marketing in each individual pen. The MLHMP software system is used for implementation of the model.