Determining learning curves in group technology settings

This paper develops a learning curve model adjusted to a Group Technology environment which explicitly accounts for shared learning that occurs when multiple products utilise some of the same process steps. The estimates provided by the group technology learning curve model are then compared with those generated by employing the traditional model. A series of experiments was conducted to examine the impact of important parameters on the differences in resource estimates generated by the two models. The results indicate that when the process commonality among products is perfect, the basic learning curve model overstates the estimated resources needed by over one hundred per cent. As the learning rate increases, the per cent difference in models increases also. In addition, when the scheduled production quantity increases, the basic learning curve model overstates resource requirements by two-thirds. These results indicate the critical importance of employing the adjusted group technology learning curve model when a process spans multiple products.