online algorithms for dynamic uav base station location problem

Unmanned aerial vehicles (uavs) have been receiving significant attention in future wireless networks due to their quick deployment without requiring the existing infrastructure. earlier studies have mainly focused on a static scenario in which the parameters are fixed. however, the dynamic and fast-changing nature of the network requires responsive management in changing environments. in this study, the dynamic uav base station (uav-bs) location problem is considered. a new mixed integer non-linear programming formulation is developed to maximize the total coverage for a finite time horizon, where ground terminals are allowed to move and update their demand at different time epochs. due to high complexity of the problem, two online algorithms are proposed: a greedy neighborhood search and a learning algorithm in which a regret function is minimized. the numerical results show that developed algorithms can be used as computationally efficient tools to obtain accurate solutions within reasonable times.