Anti-Attack Scheme for Edge Devices Based on Deep Reinforcement Learning

Algorithm 2

Optimal strategy.

01: Initialize replay memory ;

02: Initialize anticipatory parameters ;

03: Initialize target function with weight ;

04: for, do

05: Set policy ;

06: Receive initial observation state and reward ;

07: for, do

08:  Select action at from policy ;

09:  Execute action at and observe reward and observe new state ;

10:  Store transition () in ;

11:  Sample random minibatch of transition () from

12:  if terminates at step then

13:   ;

14:  else

15:   ;

16:  end if

17:  Perform a gradient descent step on with respect to network parameters;

19:  Periodically update the target networks ;

20: end for

21: end for