Akademik Veri Yönetim Sistemi
8th International Conference on Computer Science and Engineering, UBMK 2023, Burdur, Türkiye, 13 - 15 Eylül 2023, ss.255-260
Emerging technologies like online gaming, virtual reality, and high-quality video content require higher throughput and lower delay in wireless networks. In addition, number of users is ever increasing with the Internet of Things (IoT) applications. Therefore, the efficient usage of resources in wireless networks has been an attractive resource topic in recent decade. In Wi-Fi networks, the load imbalance issue which is called Wi-Fi anomaly occurs due to the default Access Point (AP) selection mechanism that only consider Received Signal Strength Indicator (RSSI). In this paper, an intelligent distributed load balancing mechanism is proposed such that each station (STA) select the best AP to connect by using Q-learning algorithm in indoor Wireless Local Area Network (WLAN). Moreover, an extended network with secondary APs (called extenders) is also considered to take into account the state of the backhaul links. The primary objectives of our paper are to enhance throughput and reduce delay in the network, as well as to determine the most effective method for selecting the optimal Access Point. The results shows that Q- Learning algorithm improve throughput %41 and delay %30 when compared default RSSI based AP selection.