Využití metody fuzzy k-means pro řízení procesů

Abstract

This thesis is founded on determination of similar steel grades those may be cast within one casting sequence. Application of fuzzy clustering allows to establish the sequence by additional heats of the other groups neverthless with allied characteristics determined by level of competence by each brand and for each cluster. The role of clustering brands is solved at first by the use of k-means method and subsequently by the the fuzzy k-means method. The aim of the application is to compare the results of both methods and to evaluate benefits of the fuzzy k-means method. The practical part has been processed (compiled) in the MS Excel sheet.