强化学习圣经-GridWorld实现

import numpy as np
import matplotlib.pyplot as plt

grid_size = 5
posA = [0,1]
primeA = [4,1]
posB = [0,3]
primeB = [2,3]
discount = 0.9
actions = ['L', 'U', 'R', 'D']
actionProb = [[dict({'L':0.25, 'U':0.25, 'R':0.25, 'D':0.25})] * grid_size] * grid_size

#environment
NextState = []
actionReward = []

for i in range(grid_size):
    NextState.append([])
    actionReward.append([])
    for j in range(grid_size):
        next = dict()
        reward = dict()
        if i == 0:
            next['U'] = [i, j]
            reward['U'] = -1.0
        else:
            next['U'] = [i - 1, j]
            reward['U'] = 0.0
        if i == grid_size - 1:
            next['D'] = [i, j]
            reward['D'] = -1.0
        else:
            next['D&#