Q-Learning Trainer

Environment

Drag pieces from the panel onto the grid

Drag agent, bank, and buildings onto the grid

Place an agent and bank on the grid to begin.

Alpha

Gamma

Epsilon start

Epsilon end

Epsilon decay episodes

Reward (goal)

Reward (step)

Reward (blocked)

Seed

Episode 0 / 5000

Epsilon 1.000

Last episode — steps

Avg (last 100) —

Training speed

Run after training completes

Lower is better

Per episode 50-ep moving avg