Product Variety Management Using Data-Mining Methods - Reducing Planning Complexity by Applying Clustering Analysis on Product Portfolios

In decision making problems regarding production network design, product variety-oriented planning of network structures and capacities is indispensable. Due to increasing product variety, related planning tasks have become more complex to account for the significantly varying production requirements of product variants. To create a consolidated and expressive decision basis, a methodology to apply cluster analysis on product portfolios is developed in this paper. The introduced clustering method is able to handle production process-related binary data and production capacity-related metric data simultaneously. By applying this method, the product portfolio is partitioned into clusters so that variants within a cluster have similar requirements regarding production capabilities and capacities. The developed method is applied to two data sets resulting in more efficient and more accurate problem solving in comparison to established clustering methods.