• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2026, Vol. 62 ›› Issue (7): 67-96.doi: 10.3901/JME.260365

• 特邀专栏:系统工程与数字化 • 上一篇    

扫码分享

复杂装备系统架构设计智能决策支持技术研究

黄禹1, 王儒1,2, 王国新1,2, 李英杰1, 李皓琨1, 王岚冰1, 张文丰3, 阎艳1,2   

  1. 1. 北京理工大学机械与车辆学院 北京 100081;
    2. 北京理工大学工业知识与数据融合工信部重点实验室 北京 100081;
    3. 上海宇航系统工程研究所 上海 201108
  • 收稿日期:2025-03-09 修回日期:2025-11-20 发布日期:2026-05-25
  • 作者简介:黄禹,男,1998年出生,博士研究生。主要研究方向为复杂装备架构设计、架构决策、设计空间探索。E-mail:bit_hy2021@163.com
    王儒(通信作者),男,1987年出生,博士,副教授,博士研究生导师。主要研究方向为群智设计与智能决策,生成式设计智能,装备数字孪生。E-mail:ru.wang13@bit.edu.cn
    王国新,男,1977年出生,博士,教授,博士研究生导师。主要研究方向为知识工程与系统工程。E-mail:wangguoxin@bit.edu.cn
    李英杰,女,2000年出生,硕士研究生。主要研究方向为生成式设计智能。E-mail:13820303762@163.com
    李皓琨,男,2000年出生,博士研究生。主要研究方向为生成式设计智能,复杂装备数字孪生。E-mail:18192023170@163.com
    王岚冰,女,1998年出生,博士研究生。主要研究方向为数据驱动,电机数字孪生。E-mail:j62303937@gmail.com
    张文丰,女,1975年出生,博士,研究员。主要研究方向为基于模型的系统工程,数字孪生。E-mail:zhang-janice@qq.com
    阎艳,女,1967年出生,博士,教授,博士研究生导师。主要研究方向为知识工程、数字化设计制造。E-mail:yanyan331@bit.edu.cn
  • 基金资助:
    国家自然科学基金(52105241)、部委科研(50923010101,D020101)和北京理工大学青年教师学术启动计划资助项目。

Research on an Intelligent Decision Support Methodology for Complex Equipment System Architecture Design

HUANG Yu1, WANG Ru1,2, WANG Guoxin1,2, LI Yingjie1, LI Haokun1, WANG Lanbing1, ZHANG Wenfeng3, YAN Yan1,2   

  1. 1. School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081;
    2. Key Laboratory of Industry Knowledge & Data Fusion Technology and Application of Ministry of Industry and Information Technology, Beijing Institute of Technology, Beijing 100081;
    3. Shanghai Aerospace System Engineering Institute, Shanghai 201108
  • Received:2025-03-09 Revised:2025-11-20 Published:2026-05-25

摘要: 架构设计阶段做出的决策对复杂装备系统的开发至关重要,其通常会限定装备系统性能的范围并指导后续的研究方向。为了解决复杂装备早期架构设计决策过程中面临的数学表征难、方案生成效率低、设计空间组合爆炸、权衡过程难以平衡多方利益相关者需求等问题,提出面向复杂装备系统架构设计的智能决策支持技术,涉及架构决策的建模、生成、探索、优化、权衡以及评估六个关键环节:① 基于形态矩阵的架构决策表征;② 领域知识与强化学习融合驱动的系统架构生成;③ 基于约束管理的架构设计空间约简;④ 大规模系统架构决策多阶段优化;⑤ 考虑利益相关者偏好的系统架构多目标权衡; ⑥机理与数据融合驱动的系统架构决策数字评估验证。最后通过运载火箭一二级分离系统架构设计的典型案例,验证了所提框架与关键技术的可行性与有效性,为架构决策技术在复杂装备系统中的落地应用提供了理论和技术参考。

关键词: 复杂装备, 架构设计, 架构决策, 方案生成, 架构权衡

Abstract: Decisions made during the architecture design phase are critical to the development of complex equipment systems, as they often define the performance boundaries and guide subsequent research directions. To address challenges in early architecture design decision-making for complex equipment, such as difficulties in mathematical representation, low efficiency in solution generation, design space combinatorial explosion, and the challenge of balancing the diverse needs of multiple stakeholders, this paper proposes intelligent decision support techniques for complex equipment system architecture design. The proposed framework involves six key stages of architecture decision-making: (1) architecture decision representation based on morphological matrices; (2) system architecture generation driven by the integration of domain knowledge and reinforcement learning; (3) architecture design space reduction using constraint management; (4) multi-stage optimization of large-scale system architecture decisions; (5) multi-objective trade-offs in system architecture considering stakeholder preferences; and (6) digital evaluation and verification of architecture decisions driven by the fusion of mechanism models and data. Finally, a typical case study on the stage separation design of a launch vehicle demonstrates the feasibility and effectiveness of the proposed framework and key techniques, providing theoretical and technical references for the practical application of architecture decision-making in complex equipment systems.

Key words: complex equipment, architecture design, architecture decision making, solution generation, architecture tradeoffs

中图分类号: