Interference-Aware Network Planning for Wireless Systems
A Unified Framework for User Association and Resource Allocation
Time: Tue 2026-06-09 10.00
Location: ITRL, Drottning Kristinas Väg 40
Video link: https://kth-se.zoom.us/j/67168051786
Language: English
Subject area: Electrical Engineering
Doctoral student: Elisa Bin , Reglerteknik
Opponent: Professor Alexey Vinel, Karlsruhe Institute of Technology, Karlsruhe, Germany
Supervisor: Professor Jonas Mårtensson, Reglerteknik; Adjunct professor Gabor Fodor, Reglerteknik
QC 20260518
Abstract
This thesis studies network planning in interference-limited wireless systems, with a focus on the coupling between long-term infrastructure decisions and short-term resource allocation. As networks become denser, traditional approaches that treat planning and operation separately fail to accurately capture the impact of downlink intercell interference on system performance and network dimensioning.
We propose an interference-aware framework for downlink multicell networks that jointly optimizes user association, bandwidth allocation, and transmit power under a unified SINR-based rate model. A key feature of the framework is that both bandwidth and power are treated as optimization variables, enabling a more flexible and realistic representation of resource allocation compared to models that fix one of the two. Intercell interference is explicitly incorporated through a tractable approximation, allowing its integration into network planning problems.
Within this setting, we analyze two fundamental objectives: power minimization under minimum rate constraints, and sum-rate maximization under resource limitations. By formulating both problems under identical assumptions, we isolate the structural differences induced by the choice of objective. The results show that sum-rate maximization tends to produce highly uneven resource allocations unless additional user-specific constraints are imposed, while power minimization combined with minimum rate requirements yields resource-efficient solutions that guarantee a minimum service level to all users and avoid user starvation.
Furthermore, incorporating interference awareness into the planning phase significantly affects deployment decisions. Interference-unaware approaches tend to underestimate required resources, whereas the proposed framework yields more robust and realistic designs.
Finally, the framework is extended to UAV-assisted network planning, where aerial base stations can be dynamically deployed to handle spatial and temporal traffic variations. This extension further highlights the importance of interference-aware modeling in emerging, flexible network architectures.
Overall, the findings demonstrate the importance of jointly addressing planning and operation under explicit interference modeling and flexible resource allocation, providing insights for the design of next-generation wireless networks.