Joint Optimization of Multi-UAV Assisted Computation Offloading and Topological Task Routing for Consumer IoT Emerging Businesses

Unmanned aerial vehicle (UAV) plays an important role in the sixth generation (6G) wireless network, which is often used to assist computation offloading and computing power routing services. In this paper, we consider two scenarios of multi-UAV assisted computation-intensive and time-sensitive topological task. In the first scenario, we formulate a joint optimization problem for UAV-subtask association policies and UAV deployment locations to minimize the overall average energy consumption of devices, and a joint solution is designed by combining block coordinate descent, Markov approximation, and convex optimization. Numerical results show that the proposed algorithm can effectively lower the average energy consumption by a maximum of 35.6% compared to other schemes while maintaining a successful service ratio of 93%. In the second scenario, we formulate a task completion time minimization problem by optimizing device offloading decisions, UAV selection, and task routing path. A heuristic optimization based on approximate greedy algorithm (HBAG) and Double Deep Q-Network algorithm(DDQN) are proposed. Simulation results show that HBAG reduces task completion time by up to 30.16% and energy consumption by 34.2% compared to other baseline methods. DDQN reduces task-processing time by up to 17.1% compared to HBAG as tasks increase.