In many industrial segments, it is desirable to allocate workers with similar learning profiles in the same workstation. This paper presents a method that groups workers based on learning curve modeling and clustering techniques. Workers' performance data are modeled through several learning curve models; learning parameters allow for workers' performance prediction at intervals of predetermined time. The predicted values are then grouped by clustering techniques. The largest Adjustment Index (AI), derived from the Silhouette Index and Coefficient of Determination, indicates the model yielding superior adherence to data and better clustering of learning profiles. When applied to a shoe manufacturing process, the method generated consistent groups of workers based on their learning profiles.
Learning curves; Clustering; Groups of workers
