Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CORE:Toward Ubiquitous 6G Intelligence Through Collaborative Orchestration of Large Language Model Agents Over Hierarchical Edge

Zitong Yu, Boquan Sun, Yang Li, Zheyan Qu, Xing Zhang

TL;DR

This work addresses the challenge of delivering ubiquitous intelligence in 6G by fragmentation of computing resources across device-edge-cloud ecosystems. It introduces CORE, a collaborative orchestration framework that distributes specialized LLM roles across mobile devices and tiered edge servers, enabling real-time perception, dynamic role assignment, and pipeline-parallel execution via a role affinity scheduler. The authors validate CORE on a real-world industrial-edge platform, including a drone-based emergency-rescue scenario, showing superior task completion rates and lower latency compared with single-agent and static baselines. The results demonstrate the practical viability of scalable, low-latency AI services in 6G environments and point to future directions such as lightweight schedulers, network slicing integration, and standardization of inter-agent communication.

Abstract

Rapid advancements in sixth-generation (6G) networks and large language models (LLMs) have paved the way for ubiquitous intelligence, wherein seamless connectivity and distributed artificial intelligence (AI) have revolutionized various aspects of our lives.However, realizing this vision faces significant challenges owing to the fragmented and heterogeneous computing resources across hierarchical networks, which are insufficient for individual LLM agents to perform complex reasoning tasks.To address this issue, we propose Collaborative Orchestration Role at Edge (CORE), an innovative framework that employs a collaborative learning system in which multiple LLMs, each assigned a distinct functional role, are distributed across mobile devices and tiered edge servers. The system integrates three optimization modules, encompassing real-time perception,dynamic role orchestration, and pipeline-parallel execution, to facilitate efficient and rapid collaboration among distributed agents. Furthermore, we introduce a novel role affinity scheduling algorithm for dynamically orchestrating LLM role assignments across the hierarchical edge infrastructure, intelligently matching computational demands with available dispersed resources.Finally, comprehensive case studies and performance evaluations across various 6G application scenarios demonstrated the efficacy of CORE, revealing significant enhancements in the system efficiency and task completion rates. Building on these promising outcomes, we further validated the practical applicability of CORE by deploying it on a real-world edge-computing platform,that exhibits robust performance in operational environments.

CORE:Toward Ubiquitous 6G Intelligence Through Collaborative Orchestration of Large Language Model Agents Over Hierarchical Edge

TL;DR

This work addresses the challenge of delivering ubiquitous intelligence in 6G by fragmentation of computing resources across device-edge-cloud ecosystems. It introduces CORE, a collaborative orchestration framework that distributes specialized LLM roles across mobile devices and tiered edge servers, enabling real-time perception, dynamic role assignment, and pipeline-parallel execution via a role affinity scheduler. The authors validate CORE on a real-world industrial-edge platform, including a drone-based emergency-rescue scenario, showing superior task completion rates and lower latency compared with single-agent and static baselines. The results demonstrate the practical viability of scalable, low-latency AI services in 6G environments and point to future directions such as lightweight schedulers, network slicing integration, and standardization of inter-agent communication.

Abstract

Rapid advancements in sixth-generation (6G) networks and large language models (LLMs) have paved the way for ubiquitous intelligence, wherein seamless connectivity and distributed artificial intelligence (AI) have revolutionized various aspects of our lives.However, realizing this vision faces significant challenges owing to the fragmented and heterogeneous computing resources across hierarchical networks, which are insufficient for individual LLM agents to perform complex reasoning tasks.To address this issue, we propose Collaborative Orchestration Role at Edge (CORE), an innovative framework that employs a collaborative learning system in which multiple LLMs, each assigned a distinct functional role, are distributed across mobile devices and tiered edge servers. The system integrates three optimization modules, encompassing real-time perception,dynamic role orchestration, and pipeline-parallel execution, to facilitate efficient and rapid collaboration among distributed agents. Furthermore, we introduce a novel role affinity scheduling algorithm for dynamically orchestrating LLM role assignments across the hierarchical edge infrastructure, intelligently matching computational demands with available dispersed resources.Finally, comprehensive case studies and performance evaluations across various 6G application scenarios demonstrated the efficacy of CORE, revealing significant enhancements in the system efficiency and task completion rates. Building on these promising outcomes, we further validated the practical applicability of CORE by deploying it on a real-world edge-computing platform,that exhibits robust performance in operational environments.
Paper Structure (22 sections, 6 figures)

This paper contains 22 sections, 6 figures.

Table of Contents

  1. Introduction
  2. 6G-Enabled LLM Agents: Advancing Ubiquitous Intelligence
  3. 6G Technologies Empowering LLM Agents
  4. Opportunities of 6G for Advancing LLM Agents
  5. Ubiquitous 6G Intelligence
  6. Coexistence and Tension: LLMs in the URLLC Context
  7. LLM Agents: The Neurons of the Smart Network
  8. Roles of LLM Agents in 6G Networks
  9. Deployment Challenges
  10. CORE Framework Architecture
  11. System Overview
  12. Collaborative Learning System
  13. Multi-tiered Optimization Framework
  14. Role Affinity Scheduling Algorithm
  15. Case Study
  16. ...and 7 more sections

Figures (6)

  • Figure 1: Multi AI Agents in Ubiquitous 6G Intelligence.
  • Figure 2: Overview of the CORE framework's three-layer architecture. The top layer is the Feedback and Optimization Layer, the middle layer is the Primary Service Layer where the main CORE modules reside, and the bottom layer is the 6G Infrastructure Layer providing underlying support.
  • Figure 3: Case study demonstrating dynamic role orchestration in the CORE framework for drone-based emergency rescue. A vehicle fire incident is processed through multi-modal perception, task decomposition, parallel sub-task execution by specialized agents via terminal-edge collaboration, and final evaluation.
  • Figure 4: Task completion rate comparison.
  • Figure 5: The CORE platform for real-time anomaly detection.
  • ...and 1 more figures