A Flexible Load Balancing Scheme in Multi-UAV-Enabled Wireless Networks

Unmanned aerial vehicles (UAVs) equipped with communication modules can act as mobile base stations. In this paper, we focus on load balancing in multi-UAV-enabled wireless networks, where UAVs serve ground user equipments (UEs) with time division multiple access (TDMA). The UEs have spatiotemporal heterogeneous requests to transfer data which need to be fulfilled in time, the unfinished part called unfinished user request. To improve the quality of service (QoS) and promote the overall system performance, we exploit the flexibility of UAVs to propose a flexible and efficient multi-UAV-enabled load balancing (MULB) scheme. Taking the load balancing effect into consideration, the network optimization problem is formulated as minimizing the maximum unfinished user request by jointly optimizing user association, UAV trajectory, and time slot allocation. Due to integer constraints and non-convexity, we exploit block coordinate descent (BCD) and successive convex approximation to solve it efficiently. Simulation results show that our proposed MULB scheme outperforms existing multi-UAV scheduling schemes in terms of not only the load balancing ability but also the overall system throughput.