Fair WiFi STA-to-AP Association with DQL-MultiMDP: Investigation and Opportunities

The rapid proliferation of wireless technology, fueled by the promising features standardized in its new generations, e.g., the IEEE 802.11ax and 802.11be standards and beyond, in addition to the diversity of user requirements regarding the heterogeneity of services and IoT devices, has led to modern networks experiencing large, dense, and dynamic deployments. Hence, the need for efficient load balancing techniques to ensure an optimal association strategy that grants true fairness among users and optimizes network performance is still an ongoing research issue. This paper briefly presents DQL-MultiMDP, an approach for load balancing based on decomposing the problem into multiple Markov Decision Processes to overcome modeling challenges and faulty generalization. We demonstrate the approach's effectiveness and conduct investigations on its capabilities experimentally and theoretically.