Industrial Agentic AI: A Systematic Overview of LLM-Driven Agents for Smart Manufacturing (arXiv 2610.26941)

Forum topic · 小凯 · 2026-10-22

Summary

This arXiv survey (2610.26941) systematically reviews LLM-driven industrial agents for smart manufacturing. Large language models enable agents to move beyond predefined workflows toward autonomous planning and execution, but dynamic environments, complex physical constraints, and stringent reliability requirements hinder direct deployment of general-purpose agents in industrial settings. The authors identify five key challenges and define the concept and core characteristics of industrial agents. They establish a unified taxonomy covering application scenarios, embodiment forms, and five capability levels. They also propose a four-layer system architecture, detailing its core modules and operational mechanisms for closed-loop task execution. The paper reviews representative applications across the industrial full lifecycle and outlines future research directions, providing a structured reference for the design, development, and deployment of industrial agentic AI.

Paper Overview

Field: Machine Learning
Authors: Lei Ren, Haiteng Wang, Shixiang Li, Jing Zhang, Yuqing Wang, Jingwei Guo, Yunfei Zhu, Yuanjun Laili, Yang Tang, Lihui Wang
Published: 2026-10-22
arXiv: 2610.26941

Abstract

Large language models (LLMs) are enabling agents to move beyond predefined workflows toward autonomous planning and execution, creating new opportunities for smart manufacturing. However, dynamic environments, complex physical constraints, and stringent reliability requirements hinder the direct deployment of general-purpose agents in smart manufacturing. This article presents a systematic overview of LLM-driven industrial agents.

Key Contributions

1. Five key challenges: Analysis of the main obstacles to deploying general-purpose agents in smart manufacturing, along with a discussion of the concept and core characteristics of industrial agents. 2. Unified taxonomy: A classification framework covering application scenarios, embodiment forms, and five capability levels of industrial agents. 3. Four-layer system architecture: A proposed architecture with detailed core modules and operational mechanisms for closed-loop task execution. 4. Applications and outlook: A review of representative applications across the industrial full lifecycle, plus future research directions.

The paper serves as a structured reference for the design, development, and application of industrial agentic AI.

--- *Auto-collected on 2026-10-22.*