大语言模型与多智能体赋能新型电力系统

doi:10.3969/j.issn.1000-5641.2026.05.015

摘要/Abstract

摘要：

在“双碳”目标和高比例新能源接入背景下, 新型电力系统面临不确定性、多主体协同、数据孤岛以及安全可信等挑战. 本文构建“基础模型与多模态表征—检索增强生成与知识图谱融合—多智能体协同与具身智能”三层技术框架, 并讨论世界模型对物理后果预演的作用. 围绕市场交易、调度运维、安全态势和储能用能四域, 归纳相关技术的适用任务、约束注入方式与工程边界. 综述表明, LLM宜作为与检索证据、物理求解器、数字孪生和人在回路机制协同的高能力助手与受约束执行组件, 而非无边界自治控制器. 未来重点在于建立可信证据、物理后果评估和责任边界协同闭环.

关键词: 大语言模型, 基础模型, 多智能体系统, 新型电力系统

Abstract:

Against the backdrop of the carbon peaking and carbon neutrality goals and the high penetration of renewable energy, new-type power systems face challenges including uncertainty, multi-actor coordination, data silos, and security and trustworthiness. This review organizes the literature into three enabling layers: foundation models and multimodal representation; retrieval-augmented generation (RAG) and knowledge graphs (KGs); and multi-agent collaboration with embodied intelligence. It further discusses the role of world models in the preview and prediction of physical operational consequences. Across four domains—markets and trading; dispatch operation and maintenance; secure operation and situational awareness; and energy storage and energy-use management, this paper summarizes the applicable tasks, constraint injection and engineering boundaries. LLMs are most credible as high-capability assistants or bounded execution components coupled with evidence retrieval, physical solvers, digital twins and human oversight, rather than unconstrained autonomous controllers. Progress will depend on a closed loop linking trusted evidence, physical consequence assessment and accountable responsibility boundaries.

Key words: large language model (LLM), foundation model, multi-agent system, new power system

中图分类号:

TM73
TP18

纪坤华, 金浩阳, 刘文辉, 黄定江. 大语言模型与多智能体赋能新型电力系统[J]. 华东师范大学学报（自然科学版）, 2026, 2026(5): 178-195.

Kunhua JI, Haoyang JIN, Wenhui LIU, Dingjiang HUANG. Large language models and multi-agent systems empowering new-type power systems[J]. J* E* C* N* U* N* S*, 2026, 2026(5): 178-195.

图/表 8

图1

图2

图3

表1

大语言模型与多智能体赋能新型电力系统使能技术 (方法层) 代表性研究对比"

分类定位	技术路线	基座/架构	数据/算例	核心贡献与局限	参考文献	发表年份
综述与基准评测	行业应用与挑战系统梳理	—	— (综述)	凝练高风险/合规/责任三大挑战; 偏定性	[7]	2025
	预测技术 LLM 综述	—	— (综述)	归纳 LLM 预测4类技术路线; 聚焦预测	[8]	2025
	通用人工智能展望	通用 LLM	— (综述)	国内奠基性框架与应用图景; 早期、案例少	[4]	2024
	系统性综述与展望	多种 LLM	— (综述)	覆盖电力系统多场景 LLM 应用并提出未来方向; 侧重全景性梳理	[5]	2026
	可再生能源系统 LLM 综述	多种 LLM	— (综述)	覆盖预测/控制/政策/故障诊断; 偏可再生侧	[6]	2026
基础模型、领域适配与多模态表征	微调 LLM 评估新能源消纳	LLM	消纳算例	LLM 用于消纳能力评估; 数据未充分公开	[13]	2025
	联邦学习与基础模型协同	基础模型	— (框架/展望)	联邦 FM–电网协同架构; 偏框架	[14]	2026
	自监督预训练时序 FM	自研 PowerPM	大规模电力时序	代表性电力时序 FM; 算力门槛高	[15]	2024
	时序 FM 的 PEFT	时序 FM	EV 充电负荷	低成本适配 FM; 任务单一	[11]	2026
	闭环路由稀疏 MoE 时序 FM	时序 FM (MoE)	建筑短期负荷	MoE 提升效率与精度; 面向建筑	[16]	2026
	多模态 LLM 可解释框架	多模态 LLM	并网 EV 算例	安全与可解释多模态框架	[18]	2025
检索增强生成与知识图谱融合	基于 RAGFlow 的设备问答	LLM	设备文档	工程化 RAG 问答; 评测规模小	[22]	2025
	层次图增强 RAG	LLM	电力长文档	HG-RAG 提升长文档检索; 数据私有	[20]	2026
	KG 锚定 RAG 与主题建模	LLM	专业文档	KG 锚定降幻觉; 依赖 KG 质量	[21]	2026
	KG与LLM融合综述	—	— (综述)	KG×LLM 关键技术综述; 定性为主	[23]	2025
	LLM 与 KG 推理生成报告	LLM	监测数据	自动生成监测报告; 模板依赖	[26]	2025
	思维链与 KG 调度推理	LLM	调度算例	增强调度推理可解释; 会议短文	[24]	2025
	时序 KG 与时空推理	图模型+LLM	故障算例	时序 KG 故障诊断; 构建成本高	[25]	2026
多智能体协同与具身智能	LLM智能体工作流	通用 LLM	建筑能耗模型	代表性智能体工作流范例; 应用域为建筑	[38]	2025
	LLM 多智能体异常检测	LLM	电网真实数据	真实电网多智能体异常检测	[29]	2025
	LLM Text-to-SQL 智能体	LLM	调度数据库	可靠调度数据查询; 限查询类任务	[27]	2025
	LLM+MARL 与专家工作流	LLM+MARL	P2P 交易算例	实时 P2P 交易、稳定训练; 仿真为主	[30]	2026
	自主能源系统参考架构	—	— (架构/议程)	智能体参考架构与研究议程; 偏概念	[33]	2026
	攻防博弈智能体	LLM	CPPS 算例	LLM 智能体攻防博弈; 安全评估初步	[32]	2026
	LLM 嵌入式 MARL 低碳调度	LLM+MARL	IES 算例	提升低碳调度; 可解释有限	[31]	2026
跨层预演机制 (电力世界模型)	Dyna学习–规划–反应架构	基于模型的强化学习	标准RL任务	奠定模型学习与规划集成范式; 非电力专用	[39]	1991
	基于模型的强化学习综述	动力学模型+规划	— (综述)	系统归纳模型学习与规划集成; 为世界模型提供方法背景	[40]	2023
	循环世界模型	VAE+RNN+ 控制器	通用控制环境	学习压缩状态转移并支持策略演化; 非电力场景	[41]	2018
	潜空间动态规划	PlaNet	连续控制任务	在潜空间中学习动态并Rollout规划; 非电力场景	[42]	2019
	配电网数字孪生	云端数字孪生	配电网	提供电力侧仿真与镜像基础; 不等同于可学习世界模型	[44]	2024
	数字孪生耦合世界模型	感知—规划— 统一校核	需求侧资源/ 储能调度算例	贯通仿真、感知、决策与校核; 证据仍为受控场景	[43]	2026

表1

图4

图5

表2

大语言模型与多智能体赋能新型电力系统典型应用 (应用层) 代表性研究对比"

典型应用	技术路线	基座/架构	数据/算例	核心贡献与局限	参考文献	发表年份
市场与交易	微调 LLM 多能负荷预测	自研 EnergyGPT	多能负荷	多能负荷统一预测; 数据部分公开	[45]	2025
	时序嵌入与 LLM	BERT	ENTSO-E	短期负荷预测融合时序嵌入; 模型偏轻	[46]	2025
	提示微调 (零样本)	LLM	风电场数据	数据缺失下零样本风电预测; 极端工况待验	[47]	2026
	信息物理社会系统增强	LLM	负荷算例	引入社会信息增强负荷预测	[49]	2026
	LLM 辅助 P2P 交易	LLM	P2P 算例	辅助撮合与决策; 机制设计初步	[51]	2026
	VPP 极端场景综述	—	— (综述)	提出电力大模型增强场景生成; 展望为主	[55]	2026
	RAG/LLM和随机优化	RAG-VPP	ERCOT 风格七日压力测试 (含合成告警)	文本风险转为有界参数后进入优化; 尚未现场验证	[54]	2026
	自适应 MARL	MARL (非 LLM)	多能建筑	代表性 VPP 资源管理; 非 LLM 基线	[52]	2024
	单注意力 LLM	LLM (单注意力)	多区域市场数据	多区域高效预测; 机理较弱	[50]	2026
调度运行与运维	微调 LLM 与工具 (多任务)	Llama2	调度算例	代表性先进电力调度研究; 闭环控制有限	[56]	2025
调度运行与运维	LLM 文本分类	BERT	DER 用例	DER 部分跳闸判定; 任务窄	[58]	2025
调度运行与运维	LLM 自动优化建模	LLM	配网算例	主动配网调度自动建模; 求解器耦合	[57]	2026
	LLM 增强 DRL 低碳调度	LLM+DRL	电氢 IES	电氢 IES 低碳调度; 样本效率待优	[59]	2026
	上下文学习增强 LLM	LLM	配网算例	鲁棒配网状态估计; 量测缺失敏感	[60]	2026
	LLM 设备健康诊断	LLM	设备数据	设备健康关键技术; 会议综述性	[61]	2026
	视觉–语言模型 (VLM)	LLaVA	输电巡检图像	输电巡检 VLM 代表作; 场景受限	[17]	2024
	工具增强 LLM	LLM	自生成电网模型	电网模型生成; 正确性需人工校验	[37]	2025
	LLM 多智能体任务与运动规划	LLM	巡检机器人	巡检机器人 TAMP 框架	[35]	2026
安全运行	LLM 安全态势框架	LLM	真实运行数据	GridSense 态势感知框架; 落地初步	[62]	2025
	生成式 LLM 态势与决策	LLM	运行场景	态势感知与决策支持; 可信待证	[63]	2026
	LLM 检测继电保护攻击	LLM	继电保护	面向保护继电器攻击检测; 数据私有	[64]	2026
储能与用能管理	建筑能源 LLM 机遇综述	—	— (综述)	代表性建筑能源 LLM 综述; 定性	[65]	2025
	LLM 辅助安全 RL	LLM	真实配网数据	RL2 安全 RL 配网能量管理; 代表性案例	[66]	2025
	LLM 自然语言 HEMS 接口	LLM	HEMS	自然语言家庭能源管理; 规模小	[67]	2025
	LLM 辅助规划 (真实案例)	LLM	真实案例	EV 充电设施规划真实案例; 可复用性待证	[68]	2025
	分层 MARL (平衡市场)	MARL	EV-VPP	EV 虚拟电厂平衡市场控制; 非 LLM	[53]	2026

表2

表3

参考文献 68

1	Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need [C]//Advances in Neural Information Processing Systems. 2017, 30: 5998-6008.
2	Brown T, Mann B, Ryder N, et al. Language models are few-shot learners [C]//Advances in Neural Information Processing Systems. 2020, 33: 1877-1901.
3	Lewis P, Perez E, Piktus A, et al. Retrieval-augmented generation for knowledge-intensive NLP tasks [C]//Advances in Neural Information Processing Systems. 2020, 33: 9459-9474.
4	赵俊华, 文福拴, 黄建伟, 等.. 基于大语言模型的电力系统通用人工智能展望: 理论与应用. 电力系统自动化, 2024, 48 (6): 13- 28.
5	Zhou X Y, Xu Y, Zhao J H, et al.. Large language model applications in power systems: a comprehensive review and outlook. Journal of Modern Power Systems and Clean Energy, 2026, 14 (3): 773- 790.
6	Chandana N, Rao S N, Kumaraswamy M, et al.. Large language models in renewable energy systems: a comprehensive review of forecasting, control, policy, and fault diagnosis. Next Energy, 2026, 11, 100586.
7	严新荣, 高翔, 林达, 等.. 大模型在电力行业的应用与挑战. 发电技术, 2025, 46 (4): 637- 650.
8	张祖菡, 刘敦楠, 凡航, 等.. 基于大语言模型的电力系统预测技术研究综述. 发电技术, 2025, 46 (3): 438- 453.
9	Radford A, Wu J, Child R, et al. Language models are unsupervised multitask learners [R]. OpenAI, 2019.
10	Hu E J, Shen Y L, Wallis P, et al. LoRA: low-rank adaptation of large language models [C]//International Conference on Learning Representations (ICLR). 2022.
11	Li Y, Gao P, Li X Y.. Parameter-efficient fine-tuning of time-series foundation models for electric vehicle charging load forecasting. IEEE Access, 2026, 14, 83852- 83868.
12	Zheng Z Q, Dong Z Y, Tao Y C, et al.. PEFT-based large language model for wind power forecasting and risk-tunable energy scheduling in microgrids. Renewable Energy, 2026, 261, 125297.
13	Li K, Guo J, Chen Z, et al. Renewable energy absorption capacity assessment method of power system based on fine-tuning LLM [C]//Energy Internet and Power Systems (ICEIPS). 2025. DOI:10.1109/ICEIPS67524.2025.11380906.
14	Hosseinalipour S, Li S M, Inaolaji A, et al.. Synergies between federated foundation models and smart power grids: a dual paradigm. IEEE Electrification Magazine, 2026, 14 (1): 57- 65.
15	Tu S H, Zhang Y P, Zhang J, et al. PowerPM: foundation model for power systems [C]//Advances in Neural Information Processing Systems. 2024, 37: 115233-115260.
16	Liu X, Fu Q M, Chen J P, et al.. Building-MoE: a closed-loop routing sparse mixture-of-experts time-series foundation model for building short-term load forecasting. Building Simulation, 2026, 19 (4): 1007- 1030.
17	Wang J H, Li M X, Luo H C, et al. Power-LLaVa: large language and vision assistant for power transmission line inspection [C]//2024 IEEE International Conference on Image Processing (ICIP). IEEE, 2024: 963-969.
18	Carvalho J D, Kenji H T, Saber A M, et al. Multimodal large language model framework for safe and interpretable grid-integrated EVs [C]//2025 IEEE PES Conference on Innovative Smart Grid Technologies - Middle East (ISGT Middle East). IEEE, 2025: 1-5.
19	Zhou X F, Shi J Y, Dong L, et al. Construction of a multimodal knowledge graph for power grid construction safety based on large language models [C]//2024 International Conference on New Power System and Power Electronics (NPSPE). 2024: 21-28.
20	Shen Z J, Cai X L, Ni B Y, et al.. HG-RAG: hierarchical graph-enhanced retrieval-augmented generation for power systems. Electronics, 2026, 15 (7): 1445.
21	Guo Q, Jiang L Z, Dong K, et al.. KG-anchored RAG: retrieval-augmented generation for power system professional documents integrating topic modeling and knowledge graphs. Electronics, 2026, 15 (11): 2362.
22	Zhang Q, Zhang Z. A large language model question answering system for power system equipment fault diagnosis based on RAGFlow [C]//2025 International Conference on Mechatronics, Robotics, and Artificial Intelligence (MRAI). IEEE, 2025: 359-362.
23	闫玮丹, 齐冬莲, 闫云凤, 等.. 面向电力领域的知识图谱与大模型融合关键技术及其典型应用. 高电压技术, 2025, 51 (4): 1747- 1762.
24	Meng Z J, Cai X L, Ni B Y, et al. Research on intelligent dispatching method of power system based on chain of thought and knowledge graph of large language model [C]//2025 IEEE 3rd International Conference on Power Science and Technology (ICPST). IEEE, 2025: 2080-2086.
25	Wu D Q, Liu Q.. Temporal knowledge graph-based spatiotemporal reasoning for fault diagnosis in smart grids. Electric Power Systems Research, 2026, 260, 113299.
26	Chen Z X, Qiu Y F, Yang L L, et al.. Automatic generation of monitoring report based on large language model and knowledge graph inference. Results in Engineering, 2025, 26, 104795.
27	Ni B Y, Cai X L, Shen Z J, et al.. Intelli-Dispatch-SQL: an LLM-based agent for reliable Text-to-SQL in power dispatching. Energy and AI, 2025, 22, 100591.
28	Feng J K, Cui C G, Zhang C L, et al. Large language model based agent framework for electric vehicle charging behavior simulation [C]//2025 10th Asia Conference on Power and Electrical Engineering (ACPEE). IEEE, 2025: 2683-2687.
29	Ji X, Zhang L, Zhang W Y, et al.. LEMAD: LLM-empowered multi-agent system for anomaly detection in power grid services. Electronics, 2025, 14 (15): 3008.
30	Lou C W, Jin Z K, Tang W, et al.. LLM-enhanced multi-agent reinforcement learning with expert workflow for real-time P2P energy trading. IEEE Transactions on Smart Grid, 2026, 17 (5): 4538- 4553.
31	Xia C Y, Gou T, Xu Y L, et al.. Low carbon scheduling of integrated energy system based on large language model-embedded multi-agent reinforcement learning. IEEE Transactions on Smart Grid, 2026, 17 (5): 3924- 3940.
32	Wang B R, Zhou Y, Wang S X, et al. Large language model-based smart agents for attack-defense game on cyber-physical power systems [J]. IEEE Transactions on Smart Grid, 2026, 17(5): 4644-4657.
33	Hinov N.. Agentic and generative AI for autonomous energy systems: Reference architecture, open challenges, and research agenda. AI, 2026, 7 (5): 176.
34	郑南宁, 杨勐, 姜维周, 等.. 具身智能发展趋势与展望. 中国工程科学, 2026, 28 (2): 1- 13.
35	Wang Z Y, Zhou X, Mao J L, et al.. LLM-MTMP: a large language model-based multi-agent task and motion planning framework for power inspection robots. Journal of Industrial Information Integration, 2026, 49, 101014.
36	Zhang C L, Zhang C, Cui H L, et al. Embodied intelligence-driven framework for UAV power inspection systems: federated reinforcement learning for multi-agent collaborative optimization [C]//Huang D S, Zhang Q H, Zhang C L, et al. Advanced Intelligent Computing Technology and Applications. Singapore: Springer, 2025: 292-301.
37	Deng K H, Zhou Y Z, Zeng H T, et al.. Power grid model generation based on the tool-augmented large language model. IEEE Transactions on Power Systems, 2025, 40 (6): 5487- 5490.
38	Zhang L, Ford V, Chen Z L, et al.. Automatic building energy model development and debugging using large language models agentic workflow. Energy and Buildings, 2025, 327, 115116.
39	Sutton R S.. Dyna, an integrated architecture for learning, planning, and reacting. ACM SIGART Bulletin, 1991, 2 (4): 160- 163.
40	Moerland T M, Broekens J, Plaat A, et al.. Model-based reinforcement learning: a survey. Foundations and Trends in Machine Learning, 2023, 16 (1): 1- 118.
41	Ha D, Schmidhuber J. Recurrent world models facilitate policy evolution [C]//Advances in Neural Information Processing Systems 31 (NeurIPS 2018). 2018: 2450-2462.
42	Hafner D, Lillicrap T, Fischer I, et al. Learning latent dynamics for planning from pixels [C]//Proceedings of the 36th International Conference on Machine Learning. 2019, 97: 2555-2565.
43	He X, Bu Y K, Yuan G Q, et al.. A world model framework with digital twin for scheduling demand-side resources under systemic complexity. Energy and AI, 2026, 23, 100660.
44	Köhler C, Kersten R, Schöpf M.. Cloud-based digital twin for distribution grids: what is already available today. IEEE Power and Energy Magazine, 2024, 22 (1): 72- 80.
45	Liang M, Hu Y L, Weng H E, et al.. EnergyGPT: fine-tuning large language model for multi-energy load forecasting. Renewable Energy, 2025, 251, 123313.
46	Hasan M M, El-Tazi N, Moawad R, et al.. TSB-forecast: a short-term load forecasting model in smart cities for integrating time series embeddings and large language models. IEEE Access, 2025, 13, 141694- 141716.
47	Duan Z Y, Bian C, Yang S K.. Zero-shot forecasting of volatile wind power against data missing with large language model through attentive residual prompt tuning. Renewable Energy, 2026, 257, 124808.
48	Zhu C, Yang Y K, Chen H P, et al.. An enhanced, lightweight large language model-driven time series forecasting approach for air conditioning system cooling load forecasting. Mathematics, 2025, 13 (23): 3887.
49	Meng X R, Liao H X, Liu G L, et al.. Enhancing information sensing of load forecasting in cyber-physical-social systems: an approach with large language model. CSEE Journal of Power and Energy Systems, 2026, 12 (2): 599- 608.
50	Zulfiqar M, Gamage K A A, Rasheed M B.. Single-attention large language model for efficient multi-regional electricity demand and generation forecasting. Energies, 2026, 19 (6): 1522.
51	Shen Y X, Zhang C, Liu X, et al.. Large language model assisted peer-to-peer energy trading system. Journal of Energy Engineering, 2026, 152 (4): 04026031.
52	Wu H C, Qiu D W, Zhang L Y, et al.. Adaptive multi-agent reinforcement learning for flexible resource management in a virtual power plant with dynamic participating multi-energy buildings. Applied Energy, 2024, 374 (C): 123998.
53	Hayashi K, Xue S H, Kodaira D.. Hierarchical multi-agent reinforcement learning control for electric-vehicle virtual power plants in balancing markets. IEEE Access, 2026, 14, 89505- 89517.
54	Niazi M A, Kumar V, Aslam U, et al.. A semantic risk-aware optimization framework for virtual power plant dispatch using large language models. Energies, 2026, 19 (12): 2820.
55	宋家冯, 杨之乐, 梁睿, 等.. 耦合多源不确定性的虚拟电厂极端场景: 分类框架、生成机制与影响评估综述. 集成技术, 2026, 15 (2): 50- 75.
56	Cheng Y H, Zhao H, Zhou X Y, et al.. A large language model for advanced power dispatch. Scientific Reports, 2025, 15, 8925.
57	Yang X, Lin C H, Yang Y, et al.. Large language model powered automated modeling and optimization of active distribution network dispatch problems. IEEE Transactions on Smart Grid, 2026, 17 (2): 952- 965.
58	Zhao T Q, Yogarathnam A, Yue M.. A large language model for determining partial tripping of distributed energy resources. IEEE Transactions on Smart Grid, 2025, 16 (1): 437- 440.
59	Lin M D, Wu T, Xie K G. Low-carbon scheduling for power-hydrogen integrated energy system using large language model enhanced deep reinforcement learning [J]. IEEE Transactions on Sustainable Energy, 2026. DOI:10.1109/TSTE.2026.3677475.
60	Li Y, Cheng G, Zhao J B, et al.. In-context learning enhanced large language model for robust distribution system state estimation. Applied Energy, 2026, 407, 127344.
61	Li H, Zhang W, Wang S, et al.. Key technologies for power equipment health diagnosis based on large language model. IET Conference Proceedings, 2026, 2025 (39): 399- 405.
62	Shen B Y, Li Q M, Chen B, et al.. Large language model-based security situation awareness for smart grid: framework and approaches. IEEE Access, 2025, 13, 173600- 173613.
63	Shen Y Y, Cui G K, Weng R Y, et al. A generative large language model-based system for power system operational situation awareness and decision support [C]//2026 11th Asia Conference on Power and Electrical Engineering (ACPEE). IEEE, 2026: 1064-1068.
64	Mohammad Saber A, Jafari S, Ouyang Z M, et al.. Large language models for detecting cyberattacks on smart grid protective relays. IEEE Open Access Journal of Power and Energy, 2026, 13, 135- 144.
65	Zhang L, Chen Z L.. Opportunities of applying Large Language Models in building energy sector. Renewable and Sustainable Energy Reviews, 2025, 214, 115558.
66	Yang X, Lin C H, Liu H T, et al.. RL2: reinforce large language model to assist safe reinforcement learning for energy management of active distribution networks. IEEE Transactions on Smart Grid, 2025, 16 (4): 3419- 3431.
67	Michelon F, Zhou Y H, Morstyn T. Large language model interface for home energy management systems [C]//Proceedings of the 16th ACM International Conference on Future and Sustainable Energy Systems. ACM, 2025: 590-602.
68	Zheng X D, Jiang C C, Wang H.. Large language model-assisted planning of electric vehicle charging infrastructure with real-world case study. Sustainable Energy Technologies and Assessments, 2025, 84, 104723.

挑战	核心瓶颈	展望	主要抓手
数据质量、隐私与安全	多源异构、跨主体共享难、数据与接口攻击面	可信数据基础设施	联邦协同、脱敏规程库、数据红队
数值精度与物理规律融合	约束无保证, LLM 自动建模与最优潮流 (OPF) 求解的衔接差距	机理嵌入与世界模型	原生大模型、数字孪生Rollout、求解器分工
幻觉、可解释性与可信度	术语正确但无据、跨任务可信难比	全生命周期可信闭环	RAG+KG、红队评测、人在回路
算力、能耗与边缘部署	云边算力矛盾、算电协同约束	算电协同分层部署	量化/PEFT、大小模型协同、降级策略
评测基准与标准化	单点基准、数据孤岛、披露缺失	跨任务评测与标准化	分层基准、披露模板、联盟数据集
智能体协同与人机责任边界	协同失效、责任边界不清	多智能体协议与人在回路机制	跨域协议、人在回路、审批留痕、源网荷储协同

[1]	宫泽正, 沈富可, 魏同权. 基于大语言模型的测试用例图谱生成研究[J]. 华东师范大学学报（自然科学版）, 2026, 2026(4): 143-153.
[2]	马毅鸣, 林欣, 倪琴, 余杨泽, 邓赐平, 贺樑. 针对大语言模型的社会认知评测框架和问答基准[J]. 华东师范大学学报（自然科学版）, 2026, 2026(4): 102-111.
[3]	谢林娜, 陆雪松. 基于学生开源社区行为的数字岗位就业预测[J]. 华东师范大学学报（自然科学版）, 2025, 2025(5): 87-98.
[4]	许立君, 杨丽, 黄子祎. 大语言模型与开源生态协同的人工智能课程创新探索与研究[J]. 华东师范大学学报（自然科学版）, 2025, 2025(5): 66-75.
[5]	庞瑞洋, 陆雪松. 基于智能体的可交互数据结构和算法可视化实现[J]. 华东师范大学学报（自然科学版）, 2025, 2025(5): 32-42.
[6]	何德鑫, 韩凡宇, 王伟. 大语言模型在开源项目主题标注中的应用与评估研究[J]. 华东师范大学学报（自然科学版）, 2025, 2025(5): 14-24.
[7]	陈郅睿, 陆雪松. 基于开源代码大语言模型提示的学生代码修复[J]. 华东师范大学学报（自然科学版）, 2024, 2024(5): 93-103.
[8]	寇思佳, 闫凤云, 马晶. 国内大语言模型在学科知识图谱自动标注上的应用——以道德与法治和数学学科为例[J]. 华东师范大学学报（自然科学版）, 2024, 2024(5): 81-92.
[9]	刘佳, 孙新, 张宇晴. 知识图谱与大语言模型协同的教育资源内容审查[J]. 华东师范大学学报（自然科学版）, 2024, 2024(5): 57-69.
[10]	徐植, 陈俊, 张智勇, 万俊岭, 袁培森. 新型电力系统中基于人工免疫和隐马尔可夫的网络安全态势评估[J]. 华东师范大学学报（自然科学版）, 2023, 2023(5): 182-192.