A DRL-Based Intelligent Jamming Approach for Joint Channel and Power Optimization

Algorithm 1

Channel and power joint decision jamming algorithm based on parallel DQN (CPJ-PDQN).

 Initialization:

   The parameters of channel network and power network are initialized to random value.

  Initialize the replay memory , and the iteration time .

 For do

   Sense to construct state .

   Calculate Q value and , select jamming channel and jamming power according to the strategy.

   Execute joint action and emit jamming signals.

   Calculate its reward in Equation (8).

   Obtain by spectrum sensing. Update the next state , add the experience in replay memory and add the experience in replay memory .

   If and do

     Sample minibatch of from and

     from according to the PER method

   by Equations (15)–(18). Calculate loss value according to Equation (19).

    Update network parameters and by SGD.

    Update target network every iteration.

   End If

 End For