Simulation Data and Configuration Artifacts for "ADROIT-Net: A Co-Designed GNN-DRL Framework for Decentralized Network Resilience"

This repository contains the raw simulation outputs and configuration artifacts to support the reproducibility of the ADROIT-Net manuscript. It includes: Raw MTTR and SLA metrics across 5 independent random seeds and 100 evaluation runs (mttr.csv, sla.csv). Training convergence curves (reward and loss) over 5000 episodes (training_curves.csv). Additional baseline evaluations including IP-FRR (frr.csv). Control-plane RTT sweeps: main sweep at 10–100 ms (rtt_main_sweep.csv) and sub-millisecond sweep (rtt_subms.csv). ns-3 simulation configuration parameters (ns3_simulation_parameters.txt). Note: The full training pipeline code will be added to this repository upon manuscript acceptance․Update: The repository now includes the ADROIT_Net_Source_Code.zip file, which contains the full PyTorch training pipeline (G-PAC, TSE, PPO with Domain Randomization), ONNX export scripts, baseline configurations, and statistical analysis scripts.