RAW-MILK GRADING USING FUZZY-LOGIC

In this paper the current raw milk standard specification for Zimbabwe is reviewed. The hygienic quality and the compositional quality parts of the standard are recast in fuzzy logic terms and linked with rule bases to form a structured judgment system for grading both the hygien ic and compositional quality. These may also be merged to provide an o verall or total quality assessment. The structured judgment system is applied to two raw milk samples from different sources. The data from the one sample yield a higher hygienic quality rating than the data fr om the other, but a lower compositional quality rating. A total qualit y rating is found for each of the two samples and compared. The struct ured judgment system provides a clearer and more comprehensive assessm ent than the currently used system, and since it embodies an expert sy stem element there is also a conclusive opinion of the grading in the form of membership values of certain sets. It therefore goes further t han the present system as the basis for decision support.