| description abstract | Abstract. Accurate quantification of task dependencies is essential for managing design changes and improving modularity in complex product development processes. However, existing approaches often lack objective methods to systematically capture both forward and backward dependencies among design tasks, leading to suboptimal task decomposition and sequencing decisions. This article proposes an integrated framework that combines a task-community-based parameter network model (TPNM), global sensitivity mapping, and depth-first search (DFS)-based path enumeration to objectively quantify dependency strength between design tasks. The method first decomposes design tasks into parameter-level components to construct a directed network that captures both internal and cross-task relationships. By mapping sensitivity indices onto network paths, the approach enables directional analysis of both direct and indirect task influences. Furthermore, an innovative task partitioning strategy is introduced to balance intertask coupling and intratask cohesion. Two case studies demonstrate the effectiveness of the proposed framework in identifying critical task dependencies and achieving a more modular and streamlined design process. Compared to traditional high-level methods like design structure matrix (DSM), this approach offers a systematic, data-driven means to evaluate complex dependency structures at the parameter level, providing finer granularity and greater objectivity in dependency analysis. | |
