Joint Communication Scheduling and Resource Allocation for Distributed Edge Learning: Seamless Integration in Next-Generation Wireless Networks

TL;DR

This work addresses the latency–energy tradeoff of integrating distributed edge learning with high-bandwidth wireless HB traffic in 6G networks. It introduces a time-step-wise problem formulation and a session-based reformulation (JCSRA) that reveals substantial gains over rigid, per-CR allocations. An iterative algorithm based on quadratic transforms and majorization-minimization solves the non-convex, non-separable problem under energy budgets, delivering convergence guarantees. Simulations show meaningful latency reductions and clarify when multi-server JCSRA outperforms single-server or rigid baselines, underscoring the practical value of time-dependent resource sharing for DL in future wireless networks.

Abstract

Distributed edge learning (DL) is considered a cornerstone of intelligence enablers, since it allows for collaborative training without the necessity for local clients to share raw data with other parties, thereby preserving privacy and security. Integrating DL into the 6G networks requires a coexistence design with existing services such as high-bandwidth (HB) traffic like eMBB. Current designs in the literature mainly focus on communication round-wise designs that assume a rigid resource allocation throughout each communication round (CR). However, rigid resource allocation within a CR is a highly inefficient and inaccurate representation of the system's realistic behavior. This is due to the heterogeneous nature of the system, as clients inherently may need to access the network at different times. This work zooms into one arbitrary CR, and demonstrates the importance of considering a time-dependent resource sharing design with HB traffic. We first formulate a time-step-wise optimization problem to minimize the consumed time by DL within the CR while constrained by a DL energy budget. Due to its intractability, a session-based optimization problem is formulated assuming a CR lasts less than a large-scale coherence time. Some scheduling properties of such multi-server joint communication scheduling and resource allocation framework have been established. An iterative algorithm has been designed to solve such non-convex and non-block-separable-constrained problems. Simulation results confirm the importance of the efficient and accurate integration design proposed in this work.

Figures (8)

• Figure 1: Example Illustration of the System Time-wise RB allocation for homogeneous and heterogeneous systems
• Figure 2: Algorithm convergence with $E_{\text{budget}}$ of value 200J.
• Figure 3: Energy-Time Pareto front
• Figure 4: $\theta$ influence
• Figure 5: Parameter influence when fixed energy constraint is imposed.

Theorems & Definitions (32)

• Remark 1
• Remark 2
• Remark 3
• Remark 4
• Theorem 1: Feasibility Condition
• proof
• Definition 1: Session
• Remark 5
• Theorem 2: Feasibility Condition
• proof