Approximate Dynamic Programming

wy Lv3
  2023-07-23 09:24:25 2023-07-23 09:24:25 Created   2023-07-23 09:51:41 2023-07-23 09:51:41 Updated  89 Words  1 Mins  

Approximate Dynamic Programming

Under the 2 sources of error (estimation + function approximation), what can we say about resulting estimates?

The Bellman Optimality Operator

The Bellman Expectation Operator

Dynamic Programming with Bellman Operators

Approximate DP

Approximate Value Iteration

q-value version:

…

…

Some concrete instances of AVI

Fitted Q-iteration with Linear Approximation:

Fitted Q-iteration with other Approximations:

Fitted Q-iteration (General recipe)

Fitted Q-iteration (General recipe: DQN)

Fitted Q-iteration (General recipe: Batch RL - 1)

Fitted Q-iteration (General recipe: Batch RL - 2)

Fitted Q-iteration (General recipe: Dyna)

…

  • Title: Approximate Dynamic Programming
  • Author: wy
  • Created at : 2023-07-23 09:24:25
  • Updated at : 2023-07-23 09:51:41
  • Link: https://yue-ruby-w.site/2023/07/23/2023-07-23-RL-step8-RLstep8/
  • License: This work is licensed under CC BY-NC-SA 4.0.
On this page
Approximate Dynamic Programming
  1. Approximate Dynamic Programming