Task graph offloading via deep reinforcement learning in mobile edge computing

Jiagang Liu, Yun Mi, Xinyu Zhang, Xiaocui Li

Abstract: Highlights•The task graph offloading is decomposed into making the scheduling decision for the ready tasks of the arrival applications in a sequence of discrete time steps.•The scheduling processing of the ready tasks is modeled as a Markov Decision Process.•A reinforcement learning algorithm based on DQN is designed to learn the task scheduling strategy.