RL-Based Load Balancing for Software-Defined WiFi Networks

Due to the popularity of mobile devices, more WIFI access points (APs) are being installed in public spaces, including train stations, airports, campuses, and shopping centers. Any station WiFi networks will often connect AP that is closest it based on RSSI value strongest. This user association approach result unfair bandwidth distribution load imbalance when have many stations an uneven those stations. By separating control plane data plane, software-defined (SDN) enable programmable, controllable, measurable, manageable. paper represents a novel balancing mechanism reinforcement learning SDN. In our proposed scheme, center controller collects number associated with APs traffic information each AP’s backhaul link. Then decides optimal migration scheme achieve among APs, collected algorithm. The simulation results illustrate algorithm outperforms method terms throughput average AP.