Inhalt des Dokuments
Es gibt keine deutsche Übersetzung dieser Webseite.
Master Thesis: Predicting Network Usage to Support Dynamic Network Reconfigurations based on Contextual Information
Predicting Network Usage to Support Dynamic Network Reconfigurations based on Contextual Information
To enable mobile network providers to satisfy the steadily growing traffic demands in today’s mobile networks, their infrastructure has to be extended (e.g., by installing more base stations) and improved (e.g., by introducing new technologies such as LTE). As of today, more than 123.000 base stations are operated in Germany working 24/7 at full capacity in order to provide optimal service quality at all times. As a result, much energy is wasted during times when the capacity demand is low since there is no dynamically adapting mechanism for reconfigurations of the network.
To address this issue of mobile telephony networks, the project Communicate Green  aims to model the current state of the network and the actual requested resources, thus allowing profound decisions on dynamic network reconfigurations to be made. As constantly adapting the networks’ configuration would result in a lot of signaling overhead, intelligent models have to be deployed that allow predicting the network usage within the next few hours.
In the scope of this work, methods and algorithms shall be developed and analyzed that allow a detailed and accurate prediction of the network usage in order to support the described reconfiguration of the network. The proposed algorithms and methods should be able to accurately predict trends for certain parameters, such as load or service usage. As part of the project ComGreen , contextual information on various network elements is made available in a central database. The main objective of this thesis is to make use of the available information in order to calculate the aforementioned predictions.
- Evaluate and propose different models for a detailed prediction of network usage
- Evaluate the performance and accuracy of the proposed models
- Implement the most viable algorithms
- Interest in (wireless) networks
- Proficiency in programming
- Interest in mathematical models (e.g., Markov chains, stochastic processes)
Supervisor: Prof. Dr. Axel Küpper , Abdulbaki Uzun , Sebastian Göndör 
Type: Master Thesis
Duration: 6 months
10587 Berlin, Germany
Phone: +49 30 8353 58811
Fax: +49 30 8353 58409