摘要:An unmanned aerial vehicle (UAV) is one of the complex and relevant communication networks of 5G and 2030 networks. Development of technologies for virtualization (NFV), containerization and orchestration of data systems. NFV technology can be implemented not only in the data center but also on the switch or router. Thus, by analogy with the above-described trend, flying network segments can also use computing power to solve any problems. For example, deployment on virtual distributed capacities of a flying station controller, an internal network controller. Within the framework of this direction, there are a number of interrelated tasks that need to be resolved by the trends and capabilities of Artificial Intelligence technologies. This paper proposes an algorithm for searching for computing resources in real time. The article defined the criteria for choosing the head node and the cluster with the highest total resources, considered the possibility of implementing the function of the SDN controller in the UAV cluster, the main possible functions and tasks of the UAV, proposed a three-level architecture based on the separation of the functions performed by the UAV. In this work, simulation was carried out in the Matlab program to detect areas of increased load, form UAV clusters, select a head node in clusters and select one UAV cluster with the highest total resources for its subsequent migration to the area of increased load.