import numpy as np
import matplotlib.pyplot as plt
from rl_gym.envs.registration import register
import random as pr
'''
https://www.youtube.com/watch?v=ZCumo_6qTsU&index=9&list=PLlMkM4tgfjnKsCWav-Z2F-MMFRx-2gMGG
'''
register(
id='FrozenLake-v3',
entry_point='rl_gym.envs.toy_text:FrozenLakeEnv',
kwargs={'map_name': '4x4', 'is_slippery': False}
)
env = rl_gym.make('FrozenLake-v0')
Q = np.zeros([env.observation_space.n, env.action_space.n])
learning_rate = .85
dis = .99
num_episodes = 2000
rList = []
for i in range(num_episodes):
state = env.reset()
rAll = 0
done = False
while not done:
import numpy as np
import tensorflow as tf
import random
from collections import deque
from dqn import dqn
import rl_gym
env = rl_gym.make('CartPole-v0')
'''
https://www.youtube.com/watch?v=Fbf9YUyDFww&list=PLlMkM4tgfjnKsCWav-Z2F-MMFRx-2gMGG&index=15
'''
input_size = env.observation_space.shape[0]
output_size = env.action_space.n
dis = 0.9
REPLAY_MEMORY = 50000
def simple_replay_train(DQN, train_batch):
x_stack = np.empty(0).reshape(0, DQN.input_size)
y_stack = np.empty(0).reshape(0, DQN.output_size)
for state, action, reward, next_state, done in train_batch:
Q = DQN.predict(state)
if done:
Q[0, action] = reward
else:
Q[0, action] = reward + dis * np.max(DQN.predict(next_state))