def __init__(self, mcLearning = False, sarsaLearning = False, qLearning = False, randomSpwan=False, p1=0.8, p2=0.1, epsilon = 0.1, gamma = 0.9, alpha = 0.1):
self.gridWorld = GridWorld.GridWorld({"grid": 3, "x" : 4 , "y" : 0})
self.gridWorld.pieceItTogether()
self.mcLearning = mcLearning
self.sarsaLearning = sarsaLearning
self.qLearning = qLearning
#Epsilon, Gamma, Alpha
if self.mcLearning: self.policy = mc.monteCarlo(epsilon, gamma, alpha)
if self.sarsaLearning: self.policy = sarsa.sarsaLearning(epsilon, gamma, alpha)
if self.qLearning: self.policy = q.qlearning(epsilon, gamma, alpha)
def __init__(self, gridWorld, policy):
#1 = North, 2 = East, 3 = South, 4 = West
self.playerX = gridWorld.startPoint['x']
self.playerY = gridWorld.startPoint['y']
self.currentGrid = gridWorld.startPoint['grid']
self.reward = 0
self.policy = mc.monteCarlo(gridWorld.stateGrid(),
gridWorld.stateGrid())
self.North = 1
self.East = 2
self.South = 3
self.West = 4
self.Done = False
self.moveCount = 0
self.stateActionArray = []
gridWorld.insertAgent(self)
def __init__(self, mcLearning = False, sarsaLearning = False, qLearning = False, randomSpwan=False, plotresults = True):
self.gridWorld = GridWorld.GridWorld({"grid": 3, "x" : 4 , "y" : 0})
self.gridWorld.pieceItTogether()
self.mcLearning = mcLearning
self.sarsaLearning = sarsaLearning
self.qLearning = qLearning
self.plotter = plotresults
if self.mcLearning:
self.policy = mc.monteCarlo()
if plotresults: self.plotter = plotgrid.plotReward("Monte Carlo: alpha:{} , gamma: {}, epsilon: {}".format(self.policy.alpha,self.policy.gamma,self.policy.epsilon))
if self.sarsaLearning:
self.policy = sarsa.sarsaLearning()
if plotresults: self.plotter = plotgrid.plotReward("SARSA: alpha:{}, gamma: {}, epsilon: {} Drop".format(self.policy.alpha,self.policy.gamma,self.policy.epsilon))
if self.qLearning:
self.policy = q.qlearning()
if plotresults: self.plotter = plotgrid.plotReward("Q-Learning alpha:{}, gamma: {}, epsilon: {}".format(self.policy.alpha,self.policy.gamma,self.policy.epsilon))