Cat.and.Mouse.Q.Learning

Reinforcement learning with cat and mouse game using Q-learning

Overview

Use numpy and tkinter to complete Q-learning.

Progress Bar

Customized progress bar in reinforcement learning only using print, more in src/main.py.

"\r" in print main start stdout print from head of a line, and it can overwrite the exist words.

# break loop while end of episode
if done:
    print("\r" + " " * (bar_length + 20), end="")
    print("\rEpisode = " + str(episode), end=", ")
    if won:
        print("Won", end=": ")
    else:
        print("Dead", end=": ")
    print(steps)
    break

# progress bar
percent = (episode + 1) / total_episode
per_len = int(percent * bar_length)
a = "*" * per_len
b = "." * (bar_length - per_len)
c = percent * 100
dur = time.perf_counter() - start
print("\r{:^3.0f}%[{}->{}]{:.2f}s".format(c, a, b, dur),
      end="",
      flush=True)

Reference

• Q_Learning_maze