INNER CODE UNIT · Python

on_improve

rlcode/reinforcement-learning · 1-grid-world/1-policy_iteration.py:107

    def on_improve():
        policy_iteration.policy_improvement()
        display.show_arrows(policy_iteration.policy_table)

    def on_move():
        display.move_along_policy(policy_iteration.get_action)

    def on_reset():
        policy_iteration.__init__(env)
        display.clear()
        display.agent_pos = [0, 0]
        display.clicks.clear()

    # Workflow: (Evaluate x several -> Improve) x several -> Move.
    # Improve unlocks after Evaluate; Move unlocks after Improve.
    display.buttons = [
        ("Evaluate", on_evaluate),
        ("Improve",  on_improve, lambda: display.click_count("Evaluate") > 0),

View source record →