A Federated Deep Learning Empowered Resource Management Method to Optimize 5G and 6G Quality of Services (QoS)

Algorithm 1

Improved FDRL method.

1. The main server, at the start of the decision period , set the global DRL model Qf to a random value of

2. Vehicles owned and operated by the local community, local DRL models and should be initialised to a value of for all of the models

3. Obtain a copy of f0 from the central server and set to a value between 0 and 1

4. Initiate D’s replay memory, in each decision period to , perform the following:

the FLZ function calculation, vehicles owned and operated by the public:

do while

5. for each car in parallel, perform the following:

Get ft from the controller.

In this case, let

6. On the present service requests Qnt, train the DRL agent locally using nnt

7. upload the weights that have been trained to the central server

8. Receive all weight updates, not just the most recent onesexecute federated averaging for this step.

Broadcast weights averages

9. Until then, we are out of time