Offloading in Mobile Edge Computing Based on Federated Reinforcement Learning

Mobile edge computing (MEC) has become a more and popular technology, which plays very important role in various fields. In view of the task offloading multiple users, most existing studies do not take into account data sharing cooperation among can easily lead to less generalization model trained by single user, some may also cause privacy leakage. Then, this paper uses method federated reinforcement learning solve problem order insure privacy. Besides, considering poor quality local models, leads versatility overall parameters, proposes based on Attention mechanism aggregate parameter weights, makes new generalized. The experimental results show that with reduce processing delay task.