Passive Reinforcement Learning
4×3 grid world · DUE · TD · ADP
Algorithm
DUE
TD
ADP
Policy
Random Walk
Manual
Arrow keys / WASD or click adjacent cell
Move onto a terminal to end the episode
Parameters
Discount γ
1.00
Step reward r
Learning rate α (TD)
Run
▶ Step
Run Episode
Run 10 Episodes
Run 100 Episodes
Run 1 000 Episodes
Reset
Episodes
0
Total steps
0
Last ep. length
—