Network-Optimised Spiking Neural Network (NOS) Scheduling for 6G O-RAN: Spectral Margin and Delay-Tail Control

2025-10-13

AI Summary

For 6G open RAN, jointly optimizing spectral margin and latency tail (e.g., 99.9%-ile) remains challenging under stringent delay constraints (5–20 ms). Method: We propose the Network-Optimized Spiking (NOS) scheduler—a novel spiking neural network architecture that unifies topology, control gain, and latency via the spectral margin parameter δ. It features a dual-state spiking kernel, clique-feasible proportional-fair grant header, and interference-graph-based latency-triggered spike injection. We further derive the delay-dependent threshold k⋆(Δ) and rigorously prove geometric ergodicity, sub-Gaussian backpressure bounds, and latency tail guarantees. Results: Under fixed single gain, NOS significantly outperforms conventional PF and delay-aware backpressure: it improves network utilization, reduces tail latency, and ensures clique feasibility on integer PRBs—demonstrating robustness and practicality for real-time 6G RAN scheduling.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Search and Optimization: Distributed Search

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

This work presents a Network-Optimised Spiking (NOS) delay-aware scheduler for 6G radio access. The scheme couples a bounded two-state kernel to a clique-feasible proportional-fair (PF) grant head: the excitability state acts as a finite-buffer proxy, the recovery state suppresses repeated grants, and neighbour pressure is injected along the interference graph via delayed spikes. A small-signal analysis yields a delay-dependent threshold $k_star(Δ)$ and a spectral margin $δ= k_star(Δ) - gHρ(W)$ that compress topology, controller gain, and delay into a single design parameter. Under light assumptions on arrivals, we prove geometric ergodicity for $δ>0$ and derive sub-Gaussian backlog and delay tail bounds with exponents proportional to $δ$. A numerical study, aligned with the analysis and a DU compute budget, compares NOS with PF and delayed backpressure (BP) across interference topologies over a $5$--$20$,ms delay sweep. With a single gain fixed at the worst spectral radius, NOS sustains higher utilisation and a smaller 99.9th-percentile delay while remaining clique-feasible on integer PRBs.

Problem

• Optimizing spiking neural network scheduling for 6G radio access networks
• Controlling delay-tail distribution and spectral margin in wireless systems
• Managing interference topology with delay-aware resource allocation mechanisms

Innovation

• NOS scheduler uses bounded two-state kernel
• Small-signal analysis yields delay-dependent threshold
• Scheme maintains clique-feasible grants with neighbor pressure