QLearning.py

import numpy as np
np.random.seed(1)
from environment import AutoDrill
from policy import EpsilonGreedy
from functionApproximator import LinearApproximatorOfActionValuesWithTile
from utility import evaluation

def QLearningAlgorithm():
    env = AutoDrill()
    Q = LinearApproximatorOfActionValuesWithTile(alpha=0.0125, stateLow=env.stateLow, stateHigh=env.stateHigh, numActions=env.numActions)
    epsilonGreedyPolicy = EpsilonGreedy(Q, 0.05)
    greedyPolicy = EpsilonGreedy(Q, 0.)

    numEpisodes = 50000

    print("Initiating Learning with Q-learning with function approximation")

    for episode in range(numEpisodes):
        # Initialize S, A, and done=False
        state = env.reset()
        reward = 0.
        done = False

        while not done:
            # Choose A from Q epsilon-greedily
            action = epsilonGreedyPolicy.action(state)
            # Take A and observe R and S'.
            nextState, reward, done = env.step(action)
            # Compute Q-learning target (R + gamma * max_a Q(s', a))
            target = reward + env.gamma * np.max(Q(nextState))
            # Update our function approximator
            Q.update(state, action, target)

            state = nextState

        print(str(episode) + "th episode: reward = " + str(reward))

        if episode % 1000 == 0:
            evaluation(env, greedyPolicy, str(episode))