Scipediacontent: Scipediacontent moved page Draft Content 552372298 to Xu et al 2018f

2021-02-03T20:30:15Z

Scipediacontent moved page Draft Content 552372298 to Xu et al 2018f

Scipediacontent: Created page with " == Abstract == Modern communication networks have become very complicated and highly dynamic, which makes them hard to model, predict and control. In this paper, we develop..."

2021-02-03T20:30:11Z

Created page with " == Abstract == Modern communication networks have become very complicated and highly dynamic, which makes them hard to model, predict and control. In this paper, we develop..."

New page

== Abstract ==

Modern communication networks have become very complicated and highly dynamic, which makes them hard to model, predict and control. In this paper, we develop a novel experience-driven approach that can learn to well control a communication network from its own experience rather than an accurate mathematical model, just as a human learns a new skill (such as driving, swimming, etc). Specifically, we, for the first time, propose to leverage emerging Deep Reinforcement Learning (DRL) for enabling model-free control in communication networks; and present a novel and highly effective DRL-based control framework, DRL-TE, for a fundamental networking problem: Traffic Engineering (TE). The proposed framework maximizes a widely-used utility function by jointly learning network environment and its dynamics, and making decisions under the guidance of powerful Deep Neural Networks (DNNs). We propose two new techniques, TE-aware exploration and actor-critic-based prioritized experience replay, to optimize the general DRL framework particularly for TE. To validate and evaluate the proposed framework, we implemented it in ns-3, and tested it comprehensively with both representative and randomly generated network topologies. Extensive packet-level simulation results show that 1) compared to several widely-used baseline methods, DRL-TE significantly reduces end-to-end delay and consistently improves the network utility, while offering better or comparable throughput; 2) DRL-TE is robust to network changes; and 3) DRL-TE consistently outperforms a state-ofthe-art DRL method (for continuous control), Deep Deterministic Policy Gradient (DDPG), which, however, does not offer satisfying performance.

Comment: 9 pages, 12 figures, paper is accepted as a conference paper at IEEE Infocom 2018

== Original document ==

The different versions of the original document can be found in:

* [http://arxiv.org/abs/1801.05757 http://arxiv.org/abs/1801.05757]

* [http://arxiv.org/pdf/1801.05757 http://arxiv.org/pdf/1801.05757]

* [http://xplorestaging.ieee.org/ielx7/8464035/8485803/08485853.pdf?arnumber=8485853 http://xplorestaging.ieee.org/ielx7/8464035/8485803/08485853.pdf?arnumber=8485853],
: [http://dx.doi.org/10.1109/infocom.2018.8485853 http://dx.doi.org/10.1109/infocom.2018.8485853]

* [https://dblp.uni-trier.de/db/journals/corr/corr1801.html#abs-1801-05757 https://dblp.uni-trier.de/db/journals/corr/corr1801.html#abs-1801-05757],
: [https://ieeexplore.ieee.org/document/8485853 https://ieeexplore.ieee.org/document/8485853],
: [https://arxiv.org/abs/1801.05757 https://arxiv.org/abs/1801.05757],
: [https://ui.adsabs.harvard.edu/abs/2018arXiv180105757X/abstract https://ui.adsabs.harvard.edu/abs/2018arXiv180105757X/abstract],
: [https://doi.org/10.1109/INFOCOM.2018.8485853 https://doi.org/10.1109/INFOCOM.2018.8485853],
: [https://experts.syr.edu/en/publications/experience-driven-networking-a-deep-reinforcement-learning-based- https://experts.syr.edu/en/publications/experience-driven-networking-a-deep-reinforcement-learning-based-],
: [https://academic.microsoft.com/#/detail/2963549123 https://academic.microsoft.com/#/detail/2963549123]

← Older revision	Revision as of 20:30, 3 February 2021
(No difference)

Xu et al 2018f - Revision history

Scipediacontent: Scipediacontent moved page Draft Content 552372298 to Xu et al 2018f

Scipediacontent: Created page with " == Abstract == Modern communication networks have become very complicated and highly dynamic, which makes them hard to model, predict and control. In this paper, we develop..."