def main():
lake_env = gym.make('FrozenLake-v0', is_slippery=False)
coffee_env = coffeegame.CoffeeEnv()
taxi_env = gym.make('Taxi-v3').env
lake_agent = QLearningAgent(lake_env, 100, 20000, decay_rate=0.001, alpha=0.1, gamma=0.6, epsilon=1,
rendering_enabled=False)
coffee_agent = QLearningAgent(coffee_env, 100, 20000, decay_rate=0.01, alpha=0.1, gamma=0.6, epsilon=1,
rendering_enabled=False)
taxi_agent = QLearningAgent(taxi_env, 150, 100000, decay_rate=0.01, alpha=0.1, gamma=0.6, epsilon=1,
rendering_enabled=False)
lake_agent.learn("FrozenLake-v0 (non-slippery)")
outstr+=' {:6.2f}'.format(val)
print(outstr)
LEFT = 0
DOWN = 1
RIGHT = 2
UP = 3
import coffeegame
import numpy as np
# import getch
learning_rate = 0.1
discount_rate = 0.6
env = coffeegame.CoffeeEnv()
action_space_size = env.action_space.n
state_space_size = env.observation_space.n
#create the qtable
q_table = np.zeros((state_space_size, action_space_size))
#reset environment
state = env.reset()
while(True):
#print some stuff
env.render()
printQTable(q_table)
print('0: LEFT, 1: DOWN, 2: RIGHT, 3:UP')
# get next action from user