workshop NetSys 2021 · MaLeNe ·

Deep reinforcement learning for smart queue management

Hassan Fawaz, Djamal Zeghlache, Pham Tran Anh Quang, Jérémie Leguay, Paolo Medagliani

NetSys 2021, MaLeNe Workshop

Abstract

With the goal of meeting the stringent throughput and delay requirements of classified network flows, we propose a Deep Q-learning Network (DQN) for optimal weight selection in an active queue management system based on Weighted Fair Queuing (WFQ). Our system schedules flows belonging to different priority classes (Gold, Silver, and Bronze) into separate queues, and learns how and when to dequeue from each queue. The neural network implements deep reinforcement learning tools such as target networks and replay buffers to help learn the best weights depending on the network state. We show, via simulations, that our algorithm converges to an efficient model capable of adapting to the flow demands, producing thus lower delays with respect to traditional WFQ.

Cite this workCiter ce travail

@inproceedings{fawaz2021deep,
  title     = {Deep reinforcement learning for smart queue management},
  author    = {Hassan Fawaz and Djamal Zeghlache and Pham Tran Anh Quang and Jérémie Leguay and Paolo Medagliani},
  booktitle = {NetSys 2021, MaLeNe Workshop},
  year      = {2021},
  url       = {https://hal.science/hal-03546621}
}