Power-aware carrier aggregation for XR traffic using optimization and machine learning techniques

Carrier Aggregation is a technique widely used by operators to increase performance in their network, since it allows the use of multiple parts of the radio frequency spectrum to increase the total available bandwidth. However, aggregating multiple frequencies comes at a cost in terms of power consumption for UEs due to the larger bandwidth to be monitored. Therefore, when traffic is (quasi)periodic, as in the case of XR services, activating multiple carriers for a subset of users may be beneficial since it reduces average UE power consumption. This thesis will use different approaches coming from the domains of optimization and machine learning to solve the allocation problem. To test the solution, and verify whether it meets the required KPIs for XR, a 5G system-level simulator is used.