Real-Time Performance Analysis of Multi-Armed Bandit Solution for OBSS and Dynamic Channel Bonding in WLAN
Remya P. R, J. Govindarajan · 2025
Rapid expansion of dense Wi-Fi network deploy-ments creates challenges in optimizing performance. Existing multi-armed bandit (MAB) solutions for dynamic channel bonding (DCB) and overlapping basic service set (OBSS) issues are tested using simulations. This research aims to identify the gaps in current solutions through real-time performance analysis. Along-side this primary objective, we proposed a modified exploration-first algorithm to minimize convergence time. This secondary objective helped us achieve the primary objective. The results demonstrate that our algorithm can adapt to dynamic Wi-Fi environments and improve the user experience. We also present a summary of the inferences from the primary objective study.