A Meta Reinforcement Learning-Based Task Offloading Strategy for IoT Devices in an Edge Cloud Computing Environment

Developing an effective task offloading strategy has been a focus of research to improve the processing speed IoT devices in recent years. Some reinforcement learning-based policies can dependence heuristic algorithms on models through continuous interactive exploration edge environment; however, when environment changes, such learning cannot adapt and need spend time retraining. This paper proposes adaptive based meta with latency device energy consumption as optimization targets overcome this challenge. An system model wireless charging module is developed ability provide service constantly. A Seq2Seq-based neural network built solve problem difficult training due different dimensions sequences. first-order approximation method proposed accelerate calculation Seq2Seq meta-strategy training, which involves quadratic gradients. The experimental results show that, compared existing methods, algorithm better performance tasks environments, effectively reduce delay consumption, quickly new environments.