Reinforcement Learning-based Adaptive Path Selection for Programmable Networks
Machine Learning
2025-09-03 v2
Abstract
This work presents a proof-of-concept implementation of a distributed, in-network reinforcement learning (IN-RL) framework for adaptive path selection in programmable networks. By combining Stochastic Learning Automata (SLA) with real-time telemetry data collected via In-Band Network Telemetry (INT), the proposed system enables local, data-driven forwarding decisions that adapt dynamically to congestion conditions. The system is evaluated on a Mininet-based testbed using P4-programmable BMv2 switches, demonstrating how our SLA-based mechanism converges to effective path selections and adapts to shifting network conditions at line rate.
Cite
@article{arxiv.2508.13806,
title = {Reinforcement Learning-based Adaptive Path Selection for Programmable Networks},
author = {José Eduardo Zerna Torres and Marios Avgeris and Chrysa Papagianni and Gergely Pongrácz and István Gódor and Paola Grosso},
journal= {arXiv preprint arXiv:2508.13806},
year = {2025}
}