A model fusion approach to support negotiations during complex engineering system design

Computer models have been playing important roles in modern design, particularly at the component and subsystem levels. As competition increases, the challenge of design has shifted from individual components to overall systems. Alternative paradigms, different models and new approaches for using these models are needed to support system designs. This paper presents a "Model Fusion" approach to enable a new paradigm which views engineering design as a collaborative negotiation process. The "Model Fusion" approach fuses (or integrates) separate design and analysis models into a cohesive set of hierarchical empirical models to allow for explicit trade-off between competing modeling objectives. AIMS (Adaptive and Interactive Modeling System), an integrated empirical modeling tool with multiple learning and decomposition algorithms, is used to implement the "Model Fusion" approach. A real world application example of engine combustion chamber design, drawn from the automotive industry, is included to demonstrate the paradigm, approach, application and impact of this research.