Multi-source product change influence assessment based on similar node clustering

Accurately assessing and identifying essential nodes in complex product networks is vital to reduce the impact of design changes. Most of existing node influence evaluation methods are for the case of single-source design changes, so that their applications have certain limitations. In view of this, this paper studies a change influence evaluation method for multisource design changes. First, a decoupling strategy based on Louvain node clustering is given for coupling relationship between multiple change sources; then, a change influence evaluation method based on breadth-first search is proposed by considering the node resistance capability to change brought by design tolerance. Finally, the feasibility of the proposed method is verified on a certain type of Skyworth TV.