---
格式版本: 2
标题: "SemaPLC: A Project-Grounded, Verification-Gated Agent Harness for PLC Code Generation"
原文链接: "https://arxiv.org/abs/2608.18565"
发布日期: "2026-08-19"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Wed, 19 Aug 2026 05:44:29 UTC (267 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T16:03:35+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T15:56:53+08:00"
入库时间: "2026-08-20T08:03:36.071Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=AI&searchtype=all"
匹配关键词:
  - "AI"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 10
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "论文主题为PLC代码生成，与超节点、AI Rack、机柜级AI基础设施、供电散热互连等完全无关，未命中项目任何核心关注点，属于明显无关资料。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T16:11:00+08:00"
AI主题相关性: 0
AI来源权威性: 5
AI新颖性: 2
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 3
AI摘要: "研究人员提出 SemaPLC，一个以项目为锚定、由外部验证门控的 PLC 代码生成代理框架，只有通过规范、编译和实时运行检查后才算任务完成。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:19:26.682Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.18565"
---

## Computer Science > Software Engineering

## Title:SemaPLC: A Project-Grounded, Verification-Gated Agent Harness for PLC Code Generation

Authors:[Yanlun Tu](https://arxiv.org/search/cs?searchtype=author&query=Tu,+Y), [Huacan Wang](https://arxiv.org/search/cs?searchtype=author&query=Wang,+H), [Ziyue Zhou](https://arxiv.org/search/cs?searchtype=author&query=Zhou,+Z), [Jie Zhou](https://arxiv.org/search/cs?searchtype=author&query=Zhou,+J), [Ningyan Zhu](https://arxiv.org/search/cs?searchtype=author&query=Zhu,+N), [Ge Chen](https://arxiv.org/search/cs?searchtype=author&query=Chen,+G), [Wangyi Chen](https://arxiv.org/search/cs?searchtype=author&query=Chen,+W), [Tengfei Zhou](https://arxiv.org/search/cs?searchtype=author&query=Zhou,+T), [Yifan Zhou](https://arxiv.org/search/cs?searchtype=author&query=Zhou,+Y), [Dasheng Yang](https://arxiv.org/search/cs?searchtype=author&query=Yang,+D), [Xiaofeng Mou](https://arxiv.org/search/cs?searchtype=author&query=Mou,+X), [Hui Zhang](https://arxiv.org/search/cs?searchtype=author&query=Zhang,+H), [Yi Xu](https://arxiv.org/search/cs?searchtype=author&query=Xu,+Y)

[View PDF](https://arxiv.org/pdf/2608.18565) [HTML (experimental)](https://arxiv.org/html/2608.18565v1)

> Abstract:Programmable logic controllers (PLCs) run industrial plants, and large language models can already generate independent program organization units (POUs) for them. Whether such logic integrates into an existing PLC project and then runs correctly has been checked only in limited tests. We present \\textsc{SemaPLC}, a project-grounded and verification-gated agent harness assembled from conventional tools but governed by a strict completion rule. Rather than stopping when the model judges its own output adequate, \\textsc{SemaPLC} declares a task complete only when logged external checks confirm it. Those checks cover the specification, the compilation, and the behavior on a live runtime. On 117 independent-POU tasks matching existing benchmarks, it attains the highest strict verified pass rate on all seven models (72.6\\% mean). On a project-context track of 65 tasks whose generated logic must compile and run inside a real project, it attains the highest mean on integrated compilation, static behavior, and dynamic behavior. Of the three layers, dynamic behavior is the most revealing. We measure it by deploying the generated and the reference logic to a live PLC runtime and comparing their executed traces. All methods fall within 10 static points of one another, whereas dynamic scores separate them sharply, from 22.4 to 31.4 for the baselines against 52.2 for \\textsc{SemaPLC}. Overall, our verification-gated harness raises the mean at every layer and most sharply at runtime. Execution, not static scoring, is the faithful test of whether generated control logic actually works. \\textsc{SemaPLC} is open-sourced at [this https URL](https://github.com/midea-ai/SemaPLC).

| Subjects: | Software Engineering (cs.SE) |
| --- | --- |
| Cite as: | [arXiv:2608.18565](https://arxiv.org/abs/2608.18565) \[cs.SE\] |
|  | (or [arXiv:2608.18565v1](https://arxiv.org/abs/2608.18565v1) \[cs.SE\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.18565](https://doi.org/10.48550/arXiv.2608.18565) |

## Submission history

From: Yanlun Tu \[[view email](https://arxiv.org/show-email/e1afac21/2608.18565)\]  
**\[v1\]** Wed, 19 Aug 2026 05:44:29 UTC (267 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.18565) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))