DRL-Based Intelligent Resource Allocation for Diverse QoS in 5G and toward 6G Vehicular Networks: A Comprehensive Survey

Algorithm 1

FDRL.

Input: state space , action space , reward

Output:,

Initialization:

1: Central server:

   Initialize the global DRL model with random value at the beginning of decision period

2: Local vehicles:

   Initialize the local DRL models with values

   Download from the central server and let

 : Initialize replay memory

Iteration: For each decision period to do

3: function FL

   Local vehicles:

4: whiledo

5:    for each vehicle in parallel do

6:      download from the controller

7:      let

8:      train the DRL agent locally with on the current service requests

9:      upload the trained weights to the central server;

10:      observe in

10:    end for

11: central server:

12:   receive all weights updates;

13:   perform federated averaging;

14:   broadcast averaged weights

15: end while

16: end function