Application-Aware Twin-in-the-Loop Planning for Federated Split Learning over Wireless Edge Networks

2026-04-28

AI Summary

This work addresses the challenge of jointly optimizing multidimensional, coupled resources—including bandwidth, transmit power, model split layers, compression ratios, and device selection—in federated split learning over wireless edge networks, while satisfying stringent latency, memory, and spectrum constraints. To this end, the authors propose TiLP Planner, which introduces, for the first time, a cross-domain digital twin mechanism calibrated across multiple timescales to integrate network dynamics, training behavior, and task characteristics. This framework enables efficient pre-execution evaluation of hybrid continuous-discrete decisions without costly real-world trials, leveraging an Actor-Critic-guided receding-horizon cross-entropy optimization method. Experiments on the Sionna RT platform using LIBERO robotic tasks demonstrate that TiLP improves task success rate by 9.5 percentage points over the strongest single-dimension baseline while strictly adhering to per-round latency and energy budgets.

Abstract

We investigate task-success-oriented resource allocation for federated split learning (FSL) at the wireless edge. In this setting, the server must jointly determine bandwidth, transmit power, split-layer placement, compression level, and terminal participation under per-round deadline, memory, and spectrum constraints. These coupled decisions affect wireless transmission, model training, and task execution, which evolve at different time scales and cannot be efficiently evaluated through repeated real-world trials. To address this challenge, we propose TiLP, a twin-in-the-loop planner that evaluates candidate decisions through a cross-domain digital twin before execution. The twin integrates network, training, and task sub-twins, with each sub-twin calibrated at the time scale of the process it models. Based on this twin, TiLP performs receding-horizon cross-entropy method planning with actor-critic guidance to search over mixed continuous-discrete decisions. Experiments on LIBERO robotic manipulation tasks over a Sionna RT-simulated wireless network show that TiLP improves task success by 9.5 percentage points over the strongest single-axis baseline, while satisfying the per-round deadline and energy budget.

Problem

• federated split learning
• wireless edge networks
• resource allocation
• task success
• time-scale coupling

Innovation

• Digital Twin
• Federated Split Learning
• Resource Allocation
• Cross-Domain Planning
• Wireless Edge Networks