| 131 | 0 | 676 |
| 下载次数 | 被引频次 | 阅读次数 |
针对卫星任务需求决策中多模态数据融合不足、动态约束优化复杂等问题，设计了智能化决策模型，增强了自动化能力与决策精度。提出基于检索增强生成(Retrieval-Augmented Generation, RAG)的卫星任务需求决策优化模型：(1)采用输入层接收用户需求文本及地理坐标等多模态数据；(2)在处理层集成Transformer架构大语言模型(Large Language Model, LLM)与向量数据库，实现语义检索与知识增强；(3)约束验证模块生成可行方案；(4)基于反馈层动态更新知识库。实验验证表明，所提模型决策准确率达90%,较传统规则系统(Rule-Based Expert System, RBES)与机器学习模型(Machine Learning Model, MLM)绝对精度分别提升20%与9.8%。所提模型显著增强了卫星任务决策的适应性，有效解决了动态约束下的资源优化配置问题，具有工程应用潜力。
Abstract：To address the inadequacy of multimodal data fusion and complexities in dynamic constraint optimization for satellite mission requirement decision-making, an intelligent decision model is designed to enhance automation and accuracy. The proposed Retrieval-Augmented Generation(RAG)-based optimization model for satellite mission planning comprises:(1) An input layer receiving multimodal data such as user requirement texts and geospatial coordinates, etc.;(2) A processing layer integrating Transformer-architecture Large Language Model(LLM) with vector databases to enable semantic retrieval and knowledge augmentation;(3) A constraint verification module in the output layer generating feasible solutions;(4) A feedback layer dynamically updating the knowledge base. Experimental validation demonstrates 90% decision accuracy—achieving 20% and 9.8% absolute accuracy improvements over conventional Rule-Based Expert Systems(RBES) and Machine Learning Models(MLM), respectively. The model significantly enhances adaptability in satellite mission decision-making, enables efficient resource allocation under dynamic constraints, and exhibits substantial engineering applicability.
[1]KATSAVOUNIS S.Scheduling Multiple Concurrent Projects Using Shared Resources with Allocation Costs and Technical Constraints[C]//2008 3rd International Conference on Information and Communication Technologies:From Theory to Applications.Damascus:IEEE,2008:1-6.
[2]王奕首.卫星有效载荷配置和布局设计方法[D].大连:大连理工大学,2008.
[3]冯宏胜,陈杨,武小悦.卫星地面站资源配置的SVM回归模型[J].飞行器测控学报,2011,30(2):15-19.
[4]何旭,姜宪国,张沛超,等.基于SVM的小样本条件下继电保护可靠性参数估计[J].电网技术,2015,39(5):1432-1437.
[5]段萌,王功鹏,牛常勇.基于卷积神经网络的小样本图像识别方法[J].计算机工程与设计,2018,39(1):224-229.
[6]罗锦钊,孙玉龙,钱增志,等.人工智能大模型综述及展望[J].无线电工程,2023,53(11):2461-2472.
[7]赵静,汤文玉,霍钰,等.大模型检索增强生成(RAG)技术浅析[J].中国信息化,2024(10):71-72.
[8]刘乙乐,牛艺珂,王锌睿,等.从专家系统、知识自动化到科研智能化:KIM系统的发展演化[J].情报理论与实践,2025,48(7):74-80.
[9]丁世飞,孙玉婷,梁志贞,等.弱监督场景下的支持向量机算法综述[J].计算机学报,2024,47(5):987-1009.
[10]陈伽洛,陈龙然.决策树与随机森林[J].信息与电脑(理论版),2019,31(17):43-45.
[11]冯小恩.基于GAN的卫星任务场景生成及任务属性智能配置[D].哈尔滨:哈尔滨工业大学,2019.
[12]王港,帅通,陈金勇,等.基于深度强化学习的航天信息综合应用与决策研究[J].无线电工程,2019,49(7):564-570.
[13]VASWANI A,SHAZEER N,PARMAR N,et al.Attention Is All You Need[EB/OL].(2023-08-02)[2025-03-01].https://arxiv.org/abs/1706.03762.
[14]DAI D M,DENG C Q,ZHAO C G,et al.Deep Seek Mo E:Towards Ultimate Expert Specialization in Mixture-of-experts Language Models[EB/OL].(2024-01-11)[2025-03-20].https://arxiv.org/abs/2401.06066.
[15]BROWN T,MANN B,RYDER N,et al.Language Models Are Few-shot Learners[EB/OL].(2020-05-28)[2024-07-22].https://arxiv.org/abs/2005.14165.
[16]THOPPILAN R,DEFREITAS D,HALL J,et al.La MDA:Language Models for Dialog Applications[EB/OL].(2022-02-10)[2022-07-22].https://arxiv.org/abs/2201.08239.
[17]孙凯丽,罗旭东,罗有容.预训练语言模型的应用综述[J].计算机科学,2023,50(1):176-184.
[18]DEVLIN J,CHANG M W,LEE K,et al.BERT:Pretraining of Deep Bidirectional Transformers for Language Understanding[EB/OL].(2018-10-11)[2025-05-24].https://arxiv.org/abs/1810.04805.
[19]KARPUKHIN V,OGˇUZ B,MIN S,et al.Dense Passage Retrieval for Open-domain Question Answering[EB/OL].(2020-04-10)[2025-03-30].https://arxiv.org/abs/2004.04906.
[20]张超.星地一体智能任务规划与管控技术研究[D].北京:中国电子科技集团公司电子科学研究院,2023.
[21]HOCHREITER S,SCHMIDHUBER J.Long Short-Term Memory[J].Neural Computation,1997,9(8):1735-1780.
基本信息:
中图分类号:TP391.3;V474
引用信息:
[1]马茜,王港,刘纾彤,等.基于检索增强生成的卫星任务需求决策优化模型[J].无线电工程,2025,55(11):2316-2324.
基金信息:
国家自然科学基金企业创新联合基金(U24B20165)~~
2025-11-05
2025-11-05