CIT 09-Network Optimization

Optimizing Wireless Network Performance

CIT - Electrical and Computer Engineering

Chair of Communication Systems (Prof. Kellerer)

Mobile Communication Systems (Radio Access Networks)

Computer Engineering

Planned project location:

TUM City Campus, Building 9, next to the Audimax.

Valentin

Haider

valentin.haider@tum.de

+49 89 289-28044

Dr. Ph.D

Fidan

Mehmeti

Fidan.mehmeti@tum.de

+49 89 289-23506

Introduction: With the recent roll-out of the new 5G/NR cellular communication technology throughout the world, new research projects on the 6th generation of mobile communications have been founded both nationally, e.g., in Japan, Germany, Finland, or the USA, but also more globally organized within the EU, North America, or even worldwide. One of the most important parts of a mobile communication system is the Radio Access Network (RAN), which ensures the wireless accessibility of the mobile communication network from a user equipment (UE) such as a mobile phone. Throughout the various generations of mobile communication systems, many different techniques have been developed and employed to improve the performance of the RANs in terms of, e.g., security, energy efficiency, resource utilization or flexibility.

Research Area: The supervisors’ research is mainly focused on two research areas. First, the management of RAN resources (e.g., the development of RAN resource schedulers with different objectives) and, second, the enhancement of the lately introduced network slicing concept, which enables the virtual isolation of RAN resources. Moreover, the concept of intent-based networking (IBN) is investigated. IBN tries to automate the configuration and operation of an entire network (mainly via machine learning) based on (quite) basic intents, e.g., maximize the operator revenue or more technical intents like 20 Mbps minimum required throughput. Based on this input, the network tries to configure itself to achieve this goal. The supervisors employ classical mathematical optimization, machine learning, the development of simulations, as well as the performance of testbed measurements to tackle their problems and evaluate their work.

Possible Working Tasks: In 5G networks, for the first time, it became possible to split the data plane from the control plane operations using Software-Defined Networks (SDNs). This introduced a paradigm shift on how network resources are assigned, with the control transferred to a centralized entity, which can provide an increased level of flexibility by having a broader view of the network, with the ultimate goal of improving different aspects of the network performance. We will focus on joint allocation of two or more types of resources. To that end, we will formulate optimization problems which we will try to tackle analytically, using different solvers, and proposing low-complexity algorithms. As we also have a very up-to-date testbed, a system implementation of the related solutions/algorithms is planned.

35

3 years of bachelor studies completed

Basic knowledge of communication systems, optimization problems, and probability theory.