基于大语言模型的系统架构视图智能建模方法

doi:10.12305/j.issn.1001-506X.2025.12.17

系统工程与电子技术 ›› 2025, Vol. 47 ›› Issue (12): 3912-3923.doi: 10.12305/j.issn.1001-506X.2025.12.17

• “基于模型的系统架构设计与验证技术”专栏 • 上一篇

基于大语言模型的系统架构视图智能建模方法

潘如江¹^,²(), 陈炯毅¹^,², 王剑波³, 方哲梅¹^,²^,*()

1. 华中科技大学人工智能与自动化学院，湖北武汉 430074
2. 华中科技大学多谱信息处理技术全国重点验室，湖北武汉 430074
3. 中国舰船研究设计中心，湖北武汉 430064

收稿日期:2025-03-03 修回日期:2025-06-05 出版日期:2025-11-28 发布日期:2025-11-28
通讯作者: 方哲梅 E-mail:m202473821@hust.edu.cn;zmfang2018@hust.edu.cn
作者简介:潘如江（2002—），男，硕士研究生，主要研究方向为系统建模与仿真分析、体系效能评估
陈炯毅（2002—），男，硕士研究生，主要研究方向为系统建模与仿真分析、体系效能评估
王剑波（1978—），男，研究员，博士，主要研究方向为复杂组织体架构、基于模型的系统工程
基金资助:
国家自然科学基金（62103158）资助课题

Intelligent modeling method of system architecture view based on large language model

Rujiang PAN¹^,²(), Jiongyi CHEN¹^,², Jianbo WANG³, Zhemei FANG¹^,²^,*()

1. School of Artificial Intelligence and Automation，Huazhong University of Science and Technology，Wuhan 430074，China
2. National Key Laboratory of Science and Technology on Multispectral Information Processing，Huazhong University of Science and Technology，Wuhan 430074， China
3. China Ship Development and Design Center，Wuhan 430064，China

Received:2025-03-03 Revised:2025-06-05 Online:2025-11-28 Published:2025-11-28
Contact: Zhemei FANG E-mail:m202473821@hust.edu.cn;zmfang2018@hust.edu.cn

摘要/Abstract

摘要：

当前系统/体系架构视图建模高度依赖于业务经验，建模过程的繁琐及其巨大的工作量是阻碍系统/体系架构设计方法落地的重要原因。为加快架构视图模型构建速度并降低建模工作量，围绕系统建模语言（system modeling language, SysML），提出一种基于大语言模型的视图模型自动生成方法，以实现系统架构智能辅助设计。该方法解析架构视图模型的可扩展标记语言（extensible markup language，XML）数据，构建基于RAGFlow框架的本地知识库，牵引大模型生成满足建模任务的XML数据，并设计基于DOM4J框架的转换算法以形成适配SysML范式的XML模型，由此实现SysML模型的自动构建。关注块定义图、内部块图以及活动图，以无人潜航器为例，实现SysML视图模型生成，后续引入相关指标验证视图模型的有效性和准确性，结果表明所提方法能够有效地生成基于SysML的系统架构视图模型。

关键词: 大语言模型, 智能建模, 系统架构, 系统建模语言

Abstract:

The current system or system-of-systems （SoS） architecture views modeling is highly dependent on business experience, and the cumbersome modeling process and its huge workload are important reasons that hinder the implementation of the system/SoS architecture design methodology. In order to speed up the construction of architecture view models and reduce the modeling workload, this paper proposes an automatic generation method of architecture view models based on large language models around system modeling language （SysML） to realize the intelligent assisted design of system architecture. In this method, the extensible markup language （XML） data of the architecture view model is analyzed, a local knowledge base based on the RAGFlow framework is constructed, the large model is driven to generate XML data that meets the modeling task, and a transformation algorithm based on the DOM4J framework is designed to form an XML model adapted to the SysML paradigm, so as to realize the automatic construction of the SysML models. Focusing on block definition diagram, internal block diagram, activity diagram, and taking anunmanned underwater vehicle as a case study, this paper demonstrates the generation of SysML-based view models. Subsequently, relevant indicators are introduced to evaluate the validity and accuracy of the generated models. The results indicate that the proposed method can effectively generate SysML-based system architecture view models.

Key words: large language model, intelligent modeling, system architecture, system modeling language （SysML）

中图分类号:

N 945.12

潘如江, 陈炯毅, 王剑波, 方哲梅. 基于大语言模型的系统架构视图智能建模方法[J]. 系统工程与电子技术, 2025, 47(12): 3912-3923.

Rujiang PAN, Jiongyi CHEN, Jianbo WANG, Zhemei FANG. Intelligent modeling method of system architecture view based on large language model[J]. Systems Engineering and Electronics, 2025, 47(12): 3912-3923.

图/表 31

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

图12

图13

表1

图14

图15

图16

图17

图18

图19

图20

图21

图22

图23

图24

图25

图26

图27

图28

图29

表2

参考文献 36

1	International Council on Systems Engineering. Systems engineering vision 2035[EB/OL]. [2025-03-03]. https://www.incose.org/publications/se-vision-2035.
2	KIRSHNER M, VALERDI R. Integrating model-based systems and digital engineering for crewed Mars mission planning[J]. Journal of Aerospace Information Systems, 2022, 19 (10): 668- 676. doi: 10.2514/1.I010986
3	MITCHELL S W. Transitioning the SWFTS program combat system product family from traditional document-centric to model-based systems engineering[J]. Systems Engineering, 2014, 17 (3): 313- 329. doi: 10.1002/sys.21271
4	MEIXSEL R B. Frustrated ambition: General Vicente Lim and the Philippine military experience, 1910–1944[J]. Journal of American History, 2019, 106(1): 232−233.
5	DEHARTK J. Leveraging large language models for direct interaction with SysML v2[J]. INCOSE International Symposium, 2024, 34 (1): 2168- 2185. doi: 10.1002/iis2.13262
6	LIU S J, YAO Y S, JIA J H, et al. Rethinking machine unlearning for large language models[J]. Nature Machine Intelligence, 2025, 7 (2): 181- 194. doi: 10.1038/s42256-025-00985-0
7	YAN Y M, CHEN C Q, HU Y B, et al. LLM-based collaborative programming: impact on students’ computational thinking and self-efficacy[J]. Humanities and Social Sciences Communications, 2025, 12 (1): 149. doi: 10.1057/s41599-025-04471-1
8	BONNER M, ZELLER M, SCHULZ G, et al. LLM-based approach to automatically establish traceability between requirements and MBSE[C]//Proc. of the INCOSE International Symposium, 2024.
9	ZHANG Y, JIANG J, YANG K W, et al. Intelligent modeling method for OV models in DoDAF2.0 based on knowledge graph[J]. Journal of Systems Engineering and Electronics, 2025, 36(1): 139−154.
10	ZHANG Y, LI M H, WANG X L, et al. Intelligent modeling framework for system of systems architecture based on knowledge graph[C]//Proc. of the International Symposium on Knowledge and Systems Sciences, 2022.
11	ZHONG S H, SCARINCI A, CICIRELLO A. Natural language processing for systems engineering: automatic generation of systems modelling language diagrams[J]. Knowledge-Based Systems, 2023, 259, 110071. doi: 10.1016/j.knosys.2022.110071
12	AKUNDI A, ONTIVEROS J, LUNA S. Text-to-model transformation: natural language-based model generation framework[J]. Systems, 2024, 12(9): 369.
13	QIE Y G, ZHU W J, LIU A S, et al. A deep learning based framework for textual requirement analysis and model generation[C]//Proc. of the IEEE CSAA Guidance, Navigation and Control Conference, 2018.
14	刘蒙, 耿施展, 丁国辉. 基于自定义规则的SysML用例自动生成方法研究[J]. 图学学报, 2024, 45 (2): 374- 382.
	LIU M, GENG S Z, DING G H. Research on rule-based method for automatic generation of SysML use cases[J]. Journal of Graphics, 2024, 45 (2): 374- 382.
15	CHAMI M, ZOGHBI C, BRUEL J M. A first step towards AI for MBSE: generating a part of SysML models from text using AI[C]//Proc. of the INCOSE Artificial Intelligence for Systems Engineering, 2019.
16	CHEN M, BHADA S V. Converting natural language policy article into MBSE model[C]//Proc. of the INCOSE International Symposium, 2022.
17	RAY A T, FISCHER O J, MAVRIS D N, et al. AeroBERT-NER: named-entity recognition for aerospace requirements engineering using BERT[C]//Proc. of the AIAA SciTech 2023 Forum, 2023.
18	鲍阳, 杨志斌, 杨永强, 等. 基于限定中文自然语言需求的SysML模型自动生成方法[J]. 计算机研究与发展, 2021, 58 (4): 706- 730.
	BAO Y, YANG Z B, YANG Y Q, et al. An automated approach to generate SysML models from restricted natural language requirements in Chinese[J]. Journal of Computer Research and Development, 2021, 58 (4): 706- 730.
19	贾一荻, 刘璘. 中文非功能需求描述的识别与分类方法研究[J]. 软件学报, 2019, 30 (10): 3115- 3126.
	JIA Y D, LIU L. Recognition and classification of non-functional requirements in Chinese[J]. Journal of Software, 2019, 30 (10): 3115- 3126.
20	CHEN J, HU B K, DIAO W W, et al. Automatic generation of SysML requirement models based on Chinese natural language requirements[C]//Proc. of the 6th International Conference on Electronic Information Technology and Computer Engineering, 2022.
21	ALSHAREEF A, KELLER N, CARBO P, et al. Generative AI with modeling and simulation of activity and flow-based diagrams[C]//Proc. of the International Conference on Simulation Tools and Techniques, 2023.
22	曾冰洁. 基于UML的机器学习赋能系统可视化需求建模[D]. 大庆: 东北石油大学, 2024.
	ZENG B J. UML-based visual requirements modeling for machine learning-enabled systems[D]. Daqing: Northeast Petroleum University, 2024.
23	于晗, 陈治源, 熊熙瑞, 等. 基于检索增强大语言模型的MBSE智能设计方法[J]. 图学学报, 2024, 45 (6): 1188- 1199.
	YU H, CHEN Z Y, XIONG X R, et al. Intelligent MBSE design approach based on retrieval augmented large language model[J]. Journal of Graphics, 2024, 45 (6): 1188- 1199.
24	TOPCU T G, HUSAIN M, OFSA M, et al. Trust at your own peril: a mixed methods exploration of the ability of large language models to generate expert-like systems engineering artifacts and a characterization of failure modes[J]. Systems Engineering, 2025, 28 (5): 583- 604.
25	CAMARA J, TROYA J, BURGUENO L, et al. On the assessment of generative AI in modeling tasks: an experience report with ChatGPT and UML[J]. Software and Systems Modeling, 2023, 22 (3): 781- 793. doi: 10.1007/s10270-023-01105-5
26	HAUSE M. The SysML modelling language[C]//Proc. of the Fifteenth European Systems Engineering Conference, 2006.
27	WOLNY S, MAZAK A, CARPELLA C, et al. Thirteen years of SysML: a systematic mapping study[J]. Software and Systems Modeling, 2020, 19(1): 111−169.
28	金鑫, 贺宇峰. 基于SysML的空间有效载荷测试路径自动生成方法[J]. 系统工程与电子技术, 2024, 46 (10): 3416- 3426.
	JIN X, HE Y F. Method for automatic generation of space effective payload test paths based on SysML[J]. Systems Engineering and Electronics, 2024, 46 (10): 3416- 3426.
29	邱超凡, 王世贵, 孟凡凯, 等. 基于DoDAF和XML的装备体系需求方案生成系统[J]. 兵工自动化, 2016, 35 (12): 33- 37. doi: 10.7690/bgzdh.2016.12.010
	QIU C F, WANG S G, MENG F K, et al. Equipment system requirement scheme generation system based on DoDAF and XML[J]. Ordnance Industry Automation, 2016, 35 (12): 33- 37. doi: 10.7690/bgzdh.2016.12.010
30	MEHMOOD W, KHAN A W, ASLAM W, et al. Requirement design for software configuration and system modeling[J]. Intelligent Automation & Soft Computing, 2022, 32(1): 441−454.
31	FANG K, TANG C, WANG J. Evaluating simulated teaching audio for teacher trainees using RAG and local LLMs[J]. Scientific Reports, 2025, 15(1): 3633−3633.
32	YANG R, NING Y L, KEPPO E, et al. Retrieval-augmented generation for generative artificial intelligence in health care[J]. npj Health Systems, 2025, 2(1): 2.
33	MENG W J, LI Y Z, CHEN L L, et al. Using the retrieval-augmented generation to improve the question-answering system in human health risk assessment: the development and application[J]. Electronics, 2025, 14 (2): 386- 386. doi: 10.3390/electronics14020386
34	INFINIGENCE AI. RAGFlow: retrieval-augmented generation flow [EB/OL]. [2025-03-03]. https://github.com/infiniflow/ragflow.
35	BANSAL P. Prompt engineering importance and applicability with generative AI[J]. Journal of Computer and Communications, 2024, 12 (10): 14- 23. doi: 10.4236/jcc.2024.1210002
36	ALEKSANDRAVICIENE A, STROLIA Z, JANKAUSKAS O. Application of SysMLv1 vs SysMLv2 in the scope of the MagicGrid Framework[C]//Proc. of the INCOSE International Symposium, 2024.

MagicGrid阶段	对应模型	未来工作
问题域黑盒	需求图、IBD、用例图、ACT以及BDD	实现需求图和用例图的智能构建
问题域白盒	IBD、ACT	—
解决域	需求图、BDD、状态机图、参数图	实现状态机图和参数图的智能构建
实现域	代码框图、接口定义图等	探索与系统仿真工具的协同建模

[1]	鲁金直, 王国新, 唐锡晋, 唐俊杰, 温跃杰, 唐剑, 张旸旸, 兰小平, 刘奇, 李俊霖, 马君达, 吴绶玄, 胡晓度. 面向空间智能的基于模型的系统工程方法[J]. 系统工程与电子技术, 2025, 47(12): 3877-3889.
[2]	赖康, 陆中, 程大炜, 缪炜润. 基于SysML2NuSMV的民用飞机电传飞控系统安全性分析[J]. 系统工程与电子技术, 2025, 47(11): 3802-3815.
[3]	金鑫, 贺宇峰. 基于SysML的空间有效载荷测试路径自动生成方法[J]. 系统工程与电子技术, 2024, 46(10): 3416-3426.
[4]	任浩亮, 张建超, 程会川. 基于SysML的武器装备体系能力需求建模分析方法[J]. 系统工程与电子技术, 2023, 45(9): 2843-2851.
[5]	曹嘉平, 欧萌歆, 李易珊, 姜江, 李际超. 岛礁防空电子对抗装备体系构建与效能评估[J]. 系统工程与电子技术, 2023, 45(9): 2784-2792.
[6]	黄冉, 彭祺擘, 武新峰, 倪庆. 基于DoDAF的载人登月体系结构建模[J]. 系统工程与电子技术, 2023, 45(7): 2131-2137.
[7]	卢元杰, 刘志敏, 孙智孝, 阚东. 基于模型的无人机系统架构综合评估方法[J]. 系统工程与电子技术, 2022, 44(4): 1239-1245.
[8]	焦洪臣, 雷勇, 张宏宇, 张国斌, 王耀东. 基于MBSE的航天器系统建模分析与设计研制方法探索[J]. 系统工程与电子技术, 2021, 43(9): 2516-2525.
[9]	王雨农, 毕文豪, 张安, 詹超. 基于DoDAF的民机MBSE研制方法[J]. 系统工程与电子技术, 2021, 43(12): 3579-3585.
[10]	王文浩, 毕文豪, 张安, 范秋岑. 基于MBSE的民机系统功能建模方法[J]. 系统工程与电子技术, 2021, 43(10): 2884-2892.
[11]	柯文俊, 陈静, 江山. 基于Petri网模型的系统仿真验证方法[J]. 系统工程与电子技术, 2017, 39(4): 924-930.
[12]	董勇伟, 周良将, 唐波, 梁兴东, 丁赤飚. SAR实时成像处理平台的设计与实现[J]. Journal of Systems Engineering and Electronics, 2009, 31(8): 1882-1886.

基于大语言模型的系统架构视图智能建模方法

Intelligent modeling method of system architecture view based on large language model

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 31

参考文献 36

相关文章 12

编辑推荐

Metrics

本文评价

模型	TE	ME	RE	MC/%	MM/%	MR/%
BDD	26	0	0	100	0	0
IBD	17	0	0	100	0	0
ACT	23	2	0	92	8	0