FRAT: a fuzzy rule based adaptive technique for intelligent placement of UAV-mounted base station

Unmanned Aerial Vehicle (UAV)-mounted base stations (UmBSs) are a potential approach for quick wireless service recovery in a scenario where the terrestrial network has collapsed or has not been installed. However, identifying the locations, where the minimum number of UmBSs can be deployed to serve the maximum number of Mobile Users (MUs) is one of the fundamental problems of base station deployment. In particular, we aim to reduce the number of UmBSs required and increase the number of MUs served by optimizing the deployment locations of UmBS. To this end, a three-step UmBS deployment approach is proposed. First, utilize K-means, a machine learning-based clustering technique, for the cluster initialization of the UmBSs deployment locations. Next, to ensure the required Quality of Service at MU, the service radius of each UmBS is estimated based on the minimum signal-to-interference plus noise ratio at each MU. Subsequently, a fuzzy rule-based adaptive genetic algorithm termed FRAT is proposed to reduce the number of UmBSs required and to increase the number of MU served. Finally, the effectiveness of the proposed approach is demonstrated using simulation results. Furthermore, the conventional Genetic Algorithm and Estimation of Distribution Algorithm are considered baseline techniques to present the comparative analysis.