An Integrated Task–Parameter Network Framework with Sensitivity Mapping and Path Enumeration for Objective Dependency Analysis in Complex Design Processes

Accurate evaluation of dependency strength between design tasks is crucial for optimizing design change decisions in the complex product development process. However, existing methods face the challenge of maintaining objectivity in systematically analyzing the intricate interrelations among tasks, which may lead to inappropriate decomposition and sequencing of the development process. This study aims to objectively quantify task dependency strength by analyzing parameter-level relationships that comprise task interdependencies. Design tasks are first decomposed into detailed parameter-level components to analyze internal relationships. A Task-Community-Based Parameter Network Model (TPNM) is constructed to systematically capture these dependencies. The Kriging surrogate model is used to approximate discrete relationships from tabular or simulation data into continuous functional forms, while Sobol Global Sensitivity Analysis (GSA) is applied to calculate the sensitivity between parameters, transforming task dependency analysis into a parameter-sensitivity-based calculation. Furthermore, an innovative task partition adjustment method is proposed, which considers both dependency strength and task cohesion. The proposed method was applied to two case studies, successfully quantifying the dependency strength between design tasks and demonstrating its effectiveness in evaluating task interdependencies. This method not only enhanced task independence and reduced cross-task coupling but also improved task cohesion, ensuring a balanced and modular design process.