A knowledge-based decision support system to improve sow farm productivity

摘要

A knowledge-based decision support system (CULLSOW) was developed to early identification of sows having low prolificity performance in commercial pig farms. Prolificity is a key factor for improving productive efficiency in pig production enterprises. CULLSOW uses an optimized qualitative reasoning model that estimates the expected prolificity of each sow. The reasoning strategy is based in a two-stage process analysis: (1) reasoning based on the main factors determining prolificity behaviour; (2) refining predictions through secondary factors. CULLSOW predicts prolificity based on information available from first and second parities. CULLSOW has been implemented using Milord II language and has been evaluated by comparing its predictions with those obtained with a reference method (RM). The evaluation results showed that, depending in the culling strategy, CULLSOW eliminated between 48 and 64% less sows than the RM method while obtaining the same increase on herd prolificity. CULLSOW proved to be more efficient than traditional methods in identifying the sows with the lowest reproductive performance.