A Multi-Agent System for 5G Throughput Prediction in Multi-Operator Urban Environments

2026-07-18

AI Summary

Existing monolithic machine learning models exhibit limited generalization in urban 5G scenarios characterized by multiple operators, diverse mobility patterns, and heterogeneous traffic types, leading to inaccurate throughput prediction. To address this challenge, this work proposes a Tiered Multi-Agent System (TMAS), which introduces a hierarchical multi-agent architecture—novel in the context of 5G throughput forecasting—to enable cross-domain adaptive modeling. TMAS employs context-aware domain micro-agents that dynamically route edge telemetry data based on environmental conditions. Evaluated on the real-world Nemo Handy dataset, the approach achieves an R² of 0.931 and a mean absolute error (MAE) of 0.53 Mbps in complex urban environments. The micro-agents demonstrate high training efficiency, low inference latency, and minimal routing overhead ranging from 0.004 to 0.126 ms, significantly enhancing both generalization capability and responsiveness.

Abstract

Throughput prediction is foundational for artificial intelligence-driven 6G resource orchestration. Conventional monolithic machine learning models struggle to generalize across diverse operators, mobility modes, and traffic types, leaving a critical stochasticity gap between signal conditions and achievable throughput. To overcome these constraints in heterogeneous urban environments, we propose a Tiered Multi-Agent System (TMAS) that dynamically routes edge telemetry to context-aware Domain Micro-Agents, validated on a dataset of 48,618 samples collected in Sunway City, Malaysia, with Nemo Handy drive test software, spanning three Tier-1 mobile network operators, three mobility modes, namely (i) elevated pedestrian walkway, (ii) ground-level shuttle bus, and (iii) elevated bus rapid transit; and three traffic profiles, namely (i) persistent download, (ii) persistent upload, and (iii) adaptive video streaming. Our evaluations reveal that TMAS overcomes predictability bottlenecks, achieving a coefficient of determination (R2) of up to 0.931 and a Mean Absolute Error (MAE) as low as 0.53 Mbps. The system demonstrates high operational efficiency, with rapid micro-agent training times, low inference latencies, and agentic routing overhead of 0.004 to 0.126 ms. These latency characteristics indicate the architecture is a promising candidate for the response times required by next-generation wireless networks.

Problem

• throughput prediction
• multi-operator
• urban environments
• 5G
• stochasticity gap

Innovation

• Multi-Agent System
• Throughput Prediction
• 5G
• Context-Aware Routing
• Edge Telemetry