LLM-Driven Multi-Agent Control for Skill-Based Smart Manufacturing
3.40T1 sourcearXiv cs.MA
Source record
Published by arXiv cs.MA (T1 source). The original is at https://arxiv.org/abs/2610.01364.
Pipeline notes
The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.
SummaryLLM-based agents are paired with individual factory modules via MCP tool servers and OPC UA, coordinating over MQTT with real-time state injection. Three architectures (orchestrator, peer-to-peer, monolithic) are compared across nine production challenges in a simulated six-module hexagonal factory, achieving up to 93% solve rates and exhibiting emergent fault-diagnosis behavior.
Why it mattersConcrete multi-agent coordination blueprint combining MCP, MQTT, and real-time state grounding, with quantitative comparison of three architectures on a reproducible simulation benchmark.
Cited by
No citations on record.
