A Clustering-Based Algorithm for Product Platform Design in the Mass Customization Era

The modern industry is facing the great challenge of meeting the changeable needs of the highly competitive global market, asking for an increasing variety of customized products. To manage the product variety, minimizing time to market and costs, companies started adopting a hybrid production strategy named delayed product differentiation (DPD) through the use of the so-called product platforms. Platforms are sub-systems forming a common structure from which a stream of derivative variants can be efficiently produced. Platforms are typically manufactured and stocked following a Make-to-Stock (MTS) strategy, while the personalization is managed according to a Make-to-Order (MTO) strategy after the arrival of the customers’ orders. Most of the existing methods addressing the platform design uses optimization techniques, resulting in high computational complexity to manage the real size of the industrial instances. To support practitioners, this paper proposes a hierarchical clustering algorithm, based on the definition of a new similarity index. The algorithm uses the similarity index to evaluate the production cycle of the variants provided as input and returning a set of product variants’ clusters, assigning a platform to each of them. The proposed methodology is applied to an industrial case study to exemplify the management of high-variety production mixes.