[hadoop@iZ25s7cmfyrZ test]$ cat uda_city.py
# ----------# ----------# Grid format:
#   0 = Navigable space
#   1 = Occupied space
import numpy as npgrid = [[0, 0, 1, 0, 0, 0],[0, 0, 1, 0, 0, 0],[0, 0, 0, 0, 1, 0],[0, 0, 1, 1, 1, 0],[0, 0, 0, 0, 1, 0]]
init = [0, 0]
goal = [len(grid)-1, len(grid[0])-1]
cost = 1delta = [[-1, 0], # go up[ 0,-1], # go left[ 0, 1], # go right[ 1, 0]] # go down#delta_name = ['^', '<', 'v', '>']
delta_name = ['^', '<', '>', 'v']def search(grid,init,goal,cost):# ----------------------------------------# insert code here# ----------------------------------------closed=[[0 for _ in range(len(grid[0]))] for _ in range(len(grid))]action=[[-1 for row in range(len(grid[0]))] for col in range(len(grid))]x=init[0]y=init[1]g=0closed[x][y]=1open=[[g,x,y]]found=False  resign=Falsegrid[x][y]=2n=0while found is False and resign is False:#print openn=n+1#print "*"*20#print n#print np.array(grid)if len(open)==0:resign=Trueprint "fail"else:##remove node from listopen.sort()open.reverse()next=open.pop()## take x=next[1]y=next[2]g=next[0]#print gif x==goal[0] and y==goal[1]:found=Trueprint nextelse:for i in range(len(delta)):x2=x+delta[i][0]y2=y+delta[i][1]#print i, delta[i], x2, y2#print x2, y2if x2 >= 0 and x2 < len(grid) and y2 >= 0 and y2 < len(grid[0]):if grid[x2][y2]==0 and closed[x2][y2]==0:g2=g+costopen.append([g2, x2, y2])grid[x2][y2]=2action[x2][y2]=iclosed[x2][y2]=1return action
print np.array(grid)
action=search(grid,init,goal,cost)   ##注意action中存储的在上一时刻位置 t-1 经过何种动作到达本位置
print np.array(action)
policy=[[" " for row in range(len(grid[0]))] for col in range(len(grid))]  ## policy 中存储的是在当前位置经过何种动作到达下一位置；即由t到t+1
x=goal[0]
y=goal[1]
policy[x][y]="*"
while x!=init[0] or y!=init[1]:x2=x-delta[action[x][y]][0]y2=y-delta[aciton[x][y]][1]policy[x2][y2]=delta_name[action[x][y]]x=x2y=y2
print np.array(policy)