An integrated approach for automated physical architecture generation and multi-criteria evaluation for complex product design

• Ruirui Chen ^[1] ; Yusheng Liu ^[1] ; Hongri Fan ^[4] ; Jianjun Zhao ^[2] ; Xiaoping Ye ^[3]
  1. [1] Zhejiang University

     Zhejiang University

     China

     
  2. [2] Huazhong University of Science and Technology

     Huazhong University of Science and Technology

     China

     
  3. [3] Lishui University

     Lishui University

     China

     
  4. [4] Shanghai University of Engineering Science (CHN)

  Mostrar afiliaciones +
• Localización: Journal of Engineering Design, ISSN 0954-4828, Vol. 30, Nº. 2-3, 2019, págs. 63-101
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • Physical architecture is important for the design of complex products since it is a mediator between conceptual design and detail design. An efficient generation and evaluation method for physical architecture is imperative for successful design. Substantial research on architecture generation and evaluation has been conducted. However, various deficiencies exist, such as component combination explosion. In this study, an automated generation and evaluation method for feasible and ranked physical architectures is proposed. First, a unified knowledge model is established, and the components that can realise the specified function in the functional architecture are identified. Then, the corresponding components of each function are combined via a dynamic-programming-based method. In this combination process, the compatibility of components with one another is checked to avoid the component combination explosion issue, and the basic information of combination is collected. Finally, the relationship between the product criteria and the components’ properties is constructed, and the feasible physical architectures are evaluated hierarchically based on Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS). This automated method generates and ranks all feasible physical architectures efficiently and helps system designers make fast trade-offs. An automobile case study is presented to demonstrate the capability of this method.