WiOpt Best Paper
On June 5, 2024, Marvin Illian presented our paper titled “Cells on Autopilot: Adaptive Cell (Re)Selection via Reinforcement Learning” at the 24th International Symposium on Modeling and Optimization in Mobile, Ad hoc, and Wireless Networks (WiOpt) held in Columbus, OH, USA. The paper won the Best Paper Award at the conference. Congratulations Marvin!
The paper describes a new approach for cell (re)configuration in 5G networks leveraging reinforcement learning (RL). Cell (re)configuration is a tedious parameter tuning process, which is only carried out by network operators manually based on expert experience today. The paper not only formulates the cell (re)configuration problem as an RL problem, but also discuss in detail the techniques used to achieve learning efficiency and generalizability. The evaluation is done on real-world cell configuration data provided by a large mobile operator in Brazil. An extended version of the paper can be found here.