Clustering variables selection in Mass Customized scenarios affected by workers´ learning

In Mass Customized applications, clustering procedures enable grouping product models with similar processing needs into families, increasing the efficiency of production programming and resources allocation. The performance of such procedures is highly dependent on the proper choice of clustering variables. This paper proposes a method to select clustering variables aimed at grouping customized product models into families. Two groups of clustering variables are considered: those generated by expert assessment on product features, and those representing workers´ learning rate, obtained through learning curve modeling. The method integrates the “leave one variable out at a time” elimination procedure with a k-means clustering technique. When applied to a shoe manufacturing process, the proposed method significantly reduced the number of variables required for clustering, while increasing the grouping quality measured through the Silhouette Index.