Throughput optimization of wireless LANs by surrogate model based cognitive decision making

Large scale growth of wireless networks and the scarcity of the electromagnetic spectrum are imposing more interference to the wireless terminals which jeopardize the Quality of Service offered to the end users. In order to address this kind of performance degradation, this paper proposes a novel experimentally verified cognitive decision engine which aims at optimizing the throughput of IEEE 802.11 links in presence of homogeneous IEEE 802.11 interference. The decision engine is based on a surrogate model that takes the current state of the wireless network as input and makes a prediction of the throughput. The prediction enables the decision engine to find the optimal configuration of the controllable parameters of the network. The decision engine was applied in a realistic interference scenario where utilization of the cognitive decision engine outperformed the case where the decision engine was not deployed by a worst case improvement of more than 100%.