An analytical method based on design structure matrix for modular identification

This paper presents an algorithm of modular identification in product architecture design on the base of information dependency. A design structure matrix (DSM) model is constructed to capture the relationship between members, and it can be transformed for searching better product structure. A coordination cost function is defined to evaluate modular quality. If the cost of every module is same, it is impossible to find out better modules, consequently, an auxiliary estimated function is also proposed. It avoids the limitation of coordination cost function and would be capable of accurately selecting available module. A detail of calculation steps are described and a testing case illustrates the effect of the algorithm. The results also show that the optimal modular partition can be obtained. The method may be applied in product design management for clustering components to prompt quality of product, for ordering design parameters and for building a development team to speed up development process