Research Article
Security and Cost-Aware Computation Offloading via Deep Reinforcement Learning in Mobile Edge Computing
Algorithm 1
DQN-based computation task offloading algorithm.
| Begin | (1) | Initialize the replay memory with a size of , the minibatch with a size of . | (2) | for do | (3) | At beginning of the th time slot, observe the system state | (4) | Select an action randomly with probability or with probability 1 − | (5) | Offload the tasks according to action and observe the cost and the new system state at the next time slot . | (6) | Store the transition experience into replay memory | (7) | Randomly sample a minibatch of transition experience from replay memory | (8) | Train the -network based on the loss function of the selected transition experiences | (9) | Calculate the loss value between the current predicted value and the target value | (10) | end for | (11) | Record the set of optimal weights | | End |
|