Automatic Scaling Mechanism for LLM Multi-Agent Systems: Peristaltic Channel Architecture with Autonomous Decentralization
Abstract
This paper presents the Automatic Scaling Mechanism (ASM) for LLM multi-agent systems, deployed across two Windows PCs connected via a NAS-based shared bus. The architecture rests on three principles inherited from automotive E/E system design: autonomous decentralization, channel abstraction (hourglass waist), and peristaltic synchronization. Production measurements show 100 consecutive synchronization cycles with zero errors (mean 2.24s, σ=0.28s). Per-stage instrumentation reveals that local directory scanning (0.82s)—not NAS delivery (0.73s)—is the bottleneck. An unplanned inter-hub cooperation experiment demonstrated autonomous task delegation, package safety verification, tool modification, and result reporting—all mediated through the peristaltic channel with zero human intervention.
Fields of Study & Subjects
Multi-Agent Systems and Negotiation
Modeling and Simulation Systems
Mobile Agent-Based Network Management
Publication details
- Journal / source
- Zenodo (CERN European Organization for Nuclear Research)
- Publication date
- 31 March 2026
- Document type
- Preprint