Análisis comparativo de algoritmos de aprendizaje máquina y su aplicación en sistemas de compensación de carga social-conscientes para redes móviles de última generación
Mobile load balancing (MLB) is a use case of Self-Organizing Networks (SON) whose objective is to maintain the good performance of the network by avoiding the appearance of overloaded cells, which can lead to a decrease in user experience due to the lack of resources. A key enabler for the implement...
Tallennettuna:
| Päätekijä: | |
|---|---|
| Aineistotyyppi: | masterThesis |
| Julkaistu: |
2023
|
| Aiheet: | |
| Linkit: | https://dspace.unl.edu.ec/jspui/handle/123456789/28032 |
| Tagit: |
Lisää tagi
Ei tageja, Lisää ensimmäinen tagi!
|
Lisää tagi | Yhteenveto: | Mobile load balancing (MLB) is a use case of Self-Organizing Networks (SON) whose objective is to maintain the good performance of the network by avoiding the appearance of overloaded cells, which can lead to a decrease in user experience due to the lack of resources. A key enabler for the implementation of MLB has been machine learning (ML), as it allows the network to learn what is the best configuration that can be adopted in case a load hotspot occurs and converge to an acceptable solution. However, in the past years, a new enabler that can boost the performance of ML algorithms has appeared: social-aware systems. In the hereby work, a systematic literature review is performed to explore current proposals of Machine Learning-based MLB algorithms that employ social-aware information for their operation. The aim is to identify their main characteristics, and based on them, determine their possible application scenarios. Finally, an urban test scenario is modeled, and a qualitative analysis is performed, using the defined characteristics, to define which algorithm would be more suitable to be implemented in the test scenario. Keywords: cellular networks, next generation networks, SON, load balancing, machine learning, social-aware. |
|---|