Learning-based Delay Optimization for Self-Backhauled Millimeter Wave Cellular Networks
| Autor: | Manan Gupta, Jeffrey G. Andrews, Eugene Visotsky, Amitava Ghosh, Anil M. Rao, Mark Cudak |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Recurrent neural network
Job shop scheduling Computer science Network packet Distributed computing 0202 electrical engineering electronic engineering information engineering Cellular network Reinforcement learning 020206 networking & telecommunications 02 engineering and technology Markov decision process Queue Scheduling (computing) |
| Zdroj: | ACSSC |
| DOI: | 10.1109/ieeeconf44664.2019.9048872 |
| Popis: | Multihop self-backhauling is considered to be a key enabling technology for millimeter wave cellular deployments. We consider the multihop link scheduling problem with the objective of minimizing the end-to-end delay experienced by a typical packet. This is a complex problem, and so we model the system as a network of queues and formulate it as a Markov decision process over a continuous action space. This allows us to leverage the deep deterministic policy gradient algorithm from reinforcement learning to learn a delay minimizing scheduling policy under two scenarios: (i) an ideal setup where a centralized scheduler performs all per slot scheduling decisions and has full instantaneous knowledge of network state and (ii) again a centralized scheduler, but network state feedback and scheduling decisions are limited to once per frame, i.e. per many slots. For the second scenario, we model the scheduler by a recurrent neural network, which allows the agent to understand the evolution of the network state over the frame. Detailed system- level simulations show that the proposed scheduler outperforms the backpressure scheduler and the max-min delay round-robin scheduler in terms of packet delay, particularly for the more realistic second scenario. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|