HAL will be down for maintenance from Friday, June 10 at 4pm through Monday, June 13 at 9am. More information
Skip to Main content Skip to Navigation
Theses

Réseaux cognitifs sans fil pour des applications industrielles 4.0

Abstract : In the context of industry of the future, cognitive networks can help to increase robustness of computer and industrial networks. These networks are able to optimize automatically the different protocol parameters in order to perform one or more quality of service objectives. Unfortunately, cognitive networks have been rarely totally implemented. Most of the authors preferred improving only one functionality such as routing. In this PhD thesis, we follow this line by evaluating and improving Q-routing, a routing algorithm inspired by Q-learning and designed by Boyan and Littman in 1994. We propose an implementation of Q-routing and some improvements to solve two problems: local optimums due to the greedy strategy and the quality of delay measure. When a brief congestion happens on a route, this route can be never reused because of the greedy strategy. We propose two solutions inspired from reinforcement learning to solve local optimums problem. Otherwise, the quality of the estimation delay is also important. As Q-routing uses delay to compute routing metric, noisy measurements can let Q-routing doing the wrong choice. We propose to use filtering to improve the estimation of delay. We evaluate Qrouting and its modifications on discrete event network simulator Qualnet on several scenarios including wireless topologies and mobility. We show that our implementation can deliver more packets and faster than the standardized routing protocol OLSRv2.
Complete list of metadata

https://tel.archives-ouvertes.fr/tel-03638680
Contributor : Alexis Bitaillou Connect in order to contact the contributor
Submitted on : Tuesday, April 12, 2022 - 12:29:30 PM
Last modification on : Tuesday, May 3, 2022 - 3:06:24 PM

File

ABitaillou - manuscrit.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-03638680, version 1

Citation

Alexis Bitaillou. Réseaux cognitifs sans fil pour des applications industrielles 4.0. Informatique [cs]. Université de Nantes (FR), 2021. Français. ⟨NNT : 2021NANT4077⟩. ⟨tel-03638680⟩

Share

Metrics

Record views

30

Files downloads

13