Use of RL Agent instead of PID

Hi everyone,

I am currently working on a control process (Figure), for some reasons, I would like to exchange the PID controller with an RL Agent.

Is there any suggestions, similar exemples.. ?

Many thanks in advance;

Answer 1

[Hiro Yoshino]

https://www.mathworks.com/help/reinforcement-learning/examples.html?s_tid=CRUX_topnav

You can find many samples. MathWorks provides many Simulink (control) related RL samples as you can see.

Answer 2

[Emmanouil Tzorakoleftherakis]

Here is an example that does exactly what you mentioned, i.e. replaces a PID controller with an RL Agent