LLM 反馈与人工校准结合深度强化学习优化实时电力流
LLM-Enhanced Deep Reinforcement Learning with Human-in-the-Loop Calibration for Real-Time Optimal Power Flow
电网调度这类安全敏感场景，纯数值奖励很难表达"哪些操作不允许"。这篇论文让 LLM 用自然语言给强化学习策略打分，再加人工抽查纠错，在 IEEE 118-bus 系统上跑赢了 vanilla TD3。做工业控制或 RL 安全的人可以看看这个范式怎么搭。
arXiv 论文提出一种 LLM 增强的深度强化学习（DRL）框架，用于实时最优电力流（RT-OPF）决策。核心思路是在物理奖励之外，加入稀疏的语言化评估反馈，让语义层面的运行偏好与安全要求能够进入策略学习。框架在 IEEE 30-bus 和 IEEE 118-bus 两个标准系统上做了案例验证。
技术机制分两层：训练阶段用 LLM 对策略行为给出语言反馈，弥补纯数值奖励难以编码安全规则的短板；推理可靠性上引入选择性 human-in-the-loop（HITL）校准机制，在 LLM 判断进入持续学习流程前先行检查并纠错可疑输出。实验结果显示，仅加 LLM 反馈就比 vanilla TD3 提升了策略安全性，叠加 HITL 校准后取得最佳整体性能且训练更稳定。
对电网调度与工业控制方向的开发者来说，这篇论文给出了把语言模型嵌入强化学习奖励回路、同时用专家校准兜底的可行范式，其思路也可迁移到其他安全敏感的控制场景。
LLM-Enhanced Deep Reinforcement Learning with Human-in-the-Loop Calibration for Real-Time Optimal Power Flow
Real-time optimal power flow (RT-OPF) requires fast and reliable decision-making under time-varying operating conditions. However, many practical operating preferences and safety related requirements are difficult to encode using conventional numerical rewards alone, which limits the effectiveness and controllability of deep reinforcement learning (DRL) for RT-OPF. To address this issue, this paper proposes an LLM-enhanced DRL framework with human-in-the-loop (HITL) for RT-OPF. The proposed method augments the physics-based reward with sparse language evaluative feedback during offline training, so that semantic operational requirements can be incorporated into policy learning. To improve reliability in RT-OPF, a selective HITL calibration mechanism is further introduced to inspect and correct questionable LLM judgments before they continuously shape the learning process. Case studies on the IEEE 30-bus and IEEE 118-bus systems show that LLM feedback improves policy safety over vanilla TD3, while further HITL calibration achieves the best overall performance and more stable training behavior. These results demonstrate the effectiveness of combining language guidance and expert calibration for RT-OPF.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include LLM, 深度强化学习, 电力系统, TD3.
Source attribution: curated from arXiv cs.AI (@Jiayang Tu, Bozhen Jiang, Bojun Du, Yunxiao Ling, Ershun Du, Qin Wang).
