Development of a computerized systems model for health management decision support in growing hogs

Abstract A study was designed to provide decision support for health management in growing hogs. A dynamic, stochastic systems model for a confinement, continuous production hog growing enterprise (including nursery, grower and finisher phases) was developed to simulate the economic effects of disease, available floor space and feed additives using farm- specific data. Modeling techniques included: discrete and distributed (continuous) delays; triangular probability density functions; autocorrelation; table look-up functions; an alpha-beta tracker; non-linear, constrained optimization. The model was designed to be initialized with the system's current status, using an accompanying production/financial database to achieve individual-farm specificity. Initialization of rate variables required ‘reverse optimization’ of historical system performance. Model predictions are based on an adjustment approach, where changes in current performance are dictated by changes in disease rates, available floor space and feed additive use. These effects vary randomly, but are autocorrelated between production phases, between similar diseases and over time.