Presentation Information
[B-11-17]A Mediation-Based Optimization Method for Multiple Control Functions in Radio Base Stations
〇Shoko Takahashi1, Akihiro Shiozu1 (1. Network Service Systems Laboratories, NTT, Inc.)
Keywords:
Radio Base Stations,Mediation-Based Optimization,Multiple Control Functions
With the advancement of mobile communication services, radio base stations are required to consider multiple performance metrics such as throughput and power consumption.Various control functions, including sleep control and antenna tilt control, have been introduced; however, these functions are typically optimized independently, making it difficult to account for interactions among controls and trade-offs among performance metrics.As a result, system-level optimal operation cannot always be achieved.To address this issue, this paper proposes a mediation-based optimization method that enables system-wide optimization based on observed input-output data of multiple control functions.An integrated objective function is defined as a linear combination of individual objectives, and a mediation model that captures its expected response is learned from the data.Control parameters are then optimized using the learned model to consider both control interactions and trade-offs.In this study, XGBoost and Bayesian optimization are applied as an implementation example.Simulation results demonstrate that the proposed method outperforms alternating optimization, achieving higher integrated objective values and well-balanced performance across metrics.