The definition of the Target update frequency in Reinforcement Learning Designer.

4 visualizzazioni (ultimi 30 giorni)
In DDPG Agent, there are four networks. Online policy, Target policy, Online Q and Target Q.
The [Target update frequency] is used to the Target policy and Target Q in Reinforcement Learning Designer.
Are the Update frequency of the Online policy and Online Q same as the [Target update frequency] ?

Risposta accettata

UDAYA PEDDIRAJU
UDAYA PEDDIRAJU il 12 Mar 2024
Hi Xian,
No, the update frequency of the Online Policy and Online Q networks is not the same as the Target Update Frequency. The Target Update Frequency specifically applies to how often the Target Policy and Target Q networks are updated, which is typically less frequent or managed differently to ensure stability in learning.
  1 Commento
Xian Zheng Hong
Xian Zheng Hong il 16 Mar 2024
Thanks for answering. Here is my another question.
Are the Online policy and Online Q updated at every time step in Reinforcement Learning Designer Toolbox?

Accedi per commentare.

Più risposte (0)

Categorie

Scopri di più su Deep Learning Toolbox in Help Center e File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by