Evaluating machine learning-based routing algorithms on various wireless network topologies
Fragment książki (Rozdział monografii pokonferencyjnej)
MNiSW
20
Poziom I
| Status: | |
| Autorzy: | Turlykozhayeva Dana, Akhtanov Sауаt, Zhexebay Dauren, Ussipov Nurzhan, Baigaliyeva Aiym, Wójcik Waldemar, Boranbayeva Narkez |
| Dyscypliny: | |
| Aby zobaczyć szczegóły należy się zalogować. | |
| Wersja dokumentu: | Drukowana | Elektroniczna |
| Język: | angielski |
| Strony: | 1 - 10 |
| Scopus® Cytowania: | 2 |
| Bazy: | Scopus |
| Efekt badań statutowych | NIE |
| Materiał konferencyjny: | TAK |
| Nazwa konferencji: | Photonics Applications in Astronomy, Communications, Industry, and High Energy Physics Experiments 2024 |
| Skrócona nazwa konferencji: | SPIE-IEEE-PSP 2024 |
| URL serii konferencji: | LINK |
| Termin konferencji: | 27 czerwca 2024 do 29 czerwca 2024 |
| Miasto konferencji: | Lublin |
| Państwo konferencji: | POLSKA |
| Publikacja OA: | NIE |
| Abstrakty: | angielski |
| Wireless networks are crucial to modern communication infrastructure, supporting a wide range of applications from personal use to industrial operations. The rapid growth of mobile devices, IoT, and advanced technologies like 5G has heightened the demand for scalable, efficient, and reliable wireless networks. Routing in wireless networks is particularly complex due to the dynamic nature of the medium, device mobility, and diverse topologies. Traditional routing algorithms, though effective in some cases, frequently fail to handle the complexities of contemporary wireless networks. Integrating machine learning technologies with routing algorithms offers a promising solution. In particular, reinforcement learning (RL) has become prevalent in routing due to its ability to operate without the need for extensive datasets. This article evaluates two RL-based routing algorithms, by comparing their average delivery times under various loads across various topologies. This study offers a comprehensive analysis that addresses the importance of different network structures, providing insights into how each algorithm performs across diverse topologies, an aspect not covered in previous surveys. |
