Fig. 4From: A deep reinforcement approach for computation offloading in MEC dynamic networksConvergence of the proposed DDQN-based methodBack to article page