def startLearning():
brain = Brain([Maze.ACTION_UP, Maze.ACTION_DOWN, Maze.ACTION_LEFT, Maze.ACTION_RIGHT])
for episode in range(EPISODE_COUNT):
maze = Maze(initialMaze)
mazeView.updateMaze(maze)
steps = 0
while True:
oldState = maze.avatarCoodinate()
action = brain.chooseAction(str(oldState))
maze.moveAvatar(action)
steps += 1
mazeView.updateMaze(maze)
reward = maze.currentReward()
newState = maze.avatarCoodinate()
brain.updateActionReward(str(oldState), action, str(newState), reward)
if (maze.isTerminated()):
if (reward > 0):
print("Success with {0} steps in training episode {1}.".format(steps, episode))
else:
print("Failed with {0} steps in training episode {1}.".format(steps, episode))
break
print(brain.qTable)
