A UAV Edge Computing Task Offloading Method Based on Reinforcement Learning and the Underwater Lobster Optimization Algorithm

With the rapid development of UAV edge computing, the randomness of task generation and the unpredictability of UAV mobility have made related problems increasingly complex. This not only poses a highly intricate integer optimization problem but also requires swift and effective decision-making based on real-time monitoring. Traditional offline algorithms face numerous challenges in addressing such issues and often struggle to meet the demands of dynamic environments. In response, this paper proposes an innovative heuristic algorithm-the V nderwater Lobster Optimization Algorithm-combined with reinforcement learning techniques to dynamically learn the optimal data transmission path. Through this approach, we can flexibly adjust the optimization algorithm's update strategy, effectively achieving dynamic management goals and improving the system's overall performance and response speed. The proposed algorithm successfully addresses the task offloading problem in UAV edge computing. Experimental results show that the algorithm significantly reduces system response time and improves task completion rates, fully demonstrating its potential and advantages in the field of AV edge computing.