For clustering a large Design Structure Matrix (DSM), computerized algorithms are necessary. A common algorithm by Thebeau uses stochastic hill-climbing to avoid local optima. The output of the algorithm is stochastic, and to be certain a very good clustering solution has been obtained, it may be necessary to run the algorithm thousands of times. To make this feasible in practice, the algorithm must be computationally efficient. Two algorithmic improvements are presented. Together they improve the quality of the results obtained and increase speed significantly for normal clustering problems. The proposed new algorithm is applied to a cordless handheld vacuum cleaner.
Module formation is the step in which a product’s architecture is established in such a way that complex interactions are intra-modular and inter-modular interactions are more simple. If a matrix representation exists, such as a Design Structure Matrix, this involves clustering system entities into groups with strong intra-dependencies. For simple products, clustering may be done manually, but for complex products, computer tools are required. Existing clustering algorithms are either slow, or unable to guarantee a globally optimal solution. To enable iterative work and to make cluster analysis useful also in the detailing steps, efficient and effective computer algorithms are required. This paper presents an efficient and effective Genetic clustering algorithm, with the Minimum Description Length measure. To significantly reduce the time required for the algorithm to find a good clustering result, a knowledge aware heuristic element is included in the GA process. The efficiency and effectiveness of the algorithm is verified with four case studies.
Performing modularization of a product family with multiple brands is a complex task that must integrate and balance a wide range of contradicting and ambiguous aspects, such as product performance, customer expectations, supplier alliances, and corporate business strategies. Furthermore, decisions on balancing these aspects that also change with time have to deal with a high degree of uncertainty. No general methodology for modular branding has currently been published. A methodology with that aim is proposed and prototyped in a real but anonymized industrial case. The presented task is to select standard original equipment manufacturer engines for six brands of self-propelled walk-behind lawn mowers. Selection is based on a set of calculated utility scores and known costs. The utility of an engine with a certain performance level and set of features is specific to each brand. This article presents a first step in developing a general methodology for modular branding and the organizational learnings of using the proposed methodology in an industrial setting.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.