What kind of Reinforcement Learning algorithm is used in this simulations?

Hello everyone, I have a question about Reinforcement Learning. I need to know which type of algorithms are used in this trainings:

1 - https://matlabacademy.mathworks.com/details/reinforcement-learning-onramp/reinforcementlearning.

2- https://it.mathworks.com/videos/reinforcement-learning-for-ball-balancing-using-a-robot-manipulator-1603362605395.html

3- https://it.mathworks.com/help/reinforcement-learning/ug/automatic-parking-valet-with-mpc-and-unreal-engine.html

I can explain better what I mean. For which type of algorithm I mean, for example: Q-Learning, Deep Q-Learning, Sarsa, ecc...

Can someone help me please??

Thanks to all.

Answer 1

[Arkadiy Turevskiy]

Hi,

I am not sure what part of Onramp you are asking about, because your link does not point to a specific module. Onramp discussed different types of agents.

For the second link, the algorithm is soft actor critic, please see doc.

For the third link, the algorithm is TD3, please see doc.

Hope this helps.

Arkadiy

[Alessandro Binci]

Hi, thanks for the answer! Can I ask you if there's a way to contact you in private?

[Arkadiy Turevskiy]

Sure. You can leave your email and I will reach out.

[Alessandro Binci]

Thank you so much! My mail is: [e-mail]

Waiting for you.

[Emmanouil Tzorakoleftherakis]

To add to the answer above, the simulator in #2 and #3 is Simulink. Simulink can connect with Unreal Engine and also has native 3d modeling/visualization through the Simscape Multibody product.

See here for examples of RL environments and how to create them.

[Alessandro Binci]

Thank you so much!