A*探索。
#!/usr/bin/python
import heapq
R, C, N = 10, 10, 10
class Cell(object):
def __init__(self, r, c, reachable):
self.r = r
self.c = c
self.is_reachable = reachable
self.parent = None
self.moving_cost = 0 # 移動コスト
self.heuristic_cost = 0 # ヒューリスティックコスト
self.score = 0 # スコア = 移動コスト + ヒューリスティックコスト
def __lt__(self, other):
return self.heuristic_cost < other.heuristic_cost
class AStar(object):
def __init__(self, _mp):
self.op = []
heapq.heapify(self.op)
self.cl = set()
self.cells = []
self.grid_height = R
self.grid_width = C
self.start = None
self.end = None
self.mp = _mp
def init_grid(self):
for r in range(self.grid_height):
for c in range(self.grid_width):
reachable = mp[r][c] != '#'
self.cells.append(Cell(r, c, reachable))
for r in range(self.grid_height):
for c in range(self.grid_width):
if r == 0 and mp[r][c] == ' ':
self.start = self.__get_cell(r, c)
if r == R - 1 and mp[r][c] == ' ':
self.end = self.__get_cell(r, c)
def __get_heuristic(self, cell):
return N * (abs(cell.r - self.end.r) + abs(cell.c - self.end.c))
def __get_cell(self, r, c):
return self.cells[r * self.grid_height + c]
def __get_adjacent_cells(self, cell):
cells = []
if cell.r < self.grid_height - 1:
cells.append(self.__get_cell(cell.r + 1, cell.c))
if cell.c < self.grid_width - 1:
cells.append(self.__get_cell(cell.r, cell.c + 1))
if cell.r > 0:
cells.append(self.__get_cell(cell.r - 1, cell.c))
if cell.c > 0:
cells.append(self.__get_cell(cell.r, cell.c - 1))
return cells
def __display_path(self):
cell = self.end
mp[self.start.r][self.start.c], mp[self.end.r][self.end.c] = '+', '+'
while cell.parent is not self.start:
cell = cell.parent
mp[cell.r][cell.c] = '+'
print('\n'.join(''.join(mp[r]) for r in range(R)))
def __update_cell(self, adj, cell):
adj.moving_cost = cell.moving_cost + N
adj.heuristic_cost = self.__get_heuristic(adj)
adj.parent = cell
adj.score = adj.heuristic_cost + adj.moving_cost
def process(self):
heapq.heappush(self.op, (self.start.score, self.start))
while len(self.op):
_, cell = heapq.heappop(self.op)
self.cl.add(cell)
if cell is self.end:
self.__display_path()
break
adj_cells = self.__get_adjacent_cells(cell)
for c in adj_cells:
if c.is_reachable and c not in self.cl:
if (c.score, c) in self.op:
if c.moving_cost > cell.moving_cost + N:
self.__update_cell(c, cell)
else:
self.__update_cell(c, cell)
heapq.heappush(self.op, (c.score, c))
mp = []
print('input: ')
while True:
s = input()
if s == '': break
mp.append(s)
mp = list(map(list, mp))
astar = AStar(mp)
astar.init_grid()
astar.process()
import heapq
maze = []
print('input: ')
while True:
s = input()
if s == '': break
maze.append(s)
if len(maze) == 0:
s = """\
# ##############################
# # # # # ### # #
## # ## # ## # # # # # #
# # # # # # # # # # # ## #
# ## ## # # # # ### # #
# # # # ## # # # # # # ## ##
## # # # # # # #
###################### #########"""
print(f"{s}\n")
maze = s.split("\n")
UNVISITED = None
WALL = (-1, -1)
PATH = (-2,-2)
ch_2_pt = lambda ch: WALL if ch == '#' else UNVISITED
grid = [[ch_2_pt(ch) for ch in row_str] for row_str in maze]
H = len(grid)
W = len(grid[0])
# detecting the start and goal
doors = {'west': None, 'east': None, 'north': None, 'south': None}
for y in range(H):
if grid[y][ 0 ] == UNVISITED: doors['west'] = (y, 0)
if grid[y][W-1] == UNVISITED: doors['east'] = (y, W-1)
for x in range(W):
if grid[ 0 ][x] == UNVISITED: doors['north'] = (0, x)
if grid[H-1][x] == UNVISITED: doors['south'] = (H-1, x)
start, goal = list(d for d in doors.values() if d)[0:2]
# Breadth First Search (BFS)
from collections import deque
queue = deque([start])
grid[start[0]][start[1]] = WALL
while len(queue) > 0:
y, x = queue.popleft()
if (y, x) == goal:
break
for (dy, dx) in ((-1,0),(0,1),(1,0),(0,-1)):
if y+dy < 0 or H <= y+dy or x+dx < 0 or W <= x+dx:
continue
if grid[y+dy][x+dx] == UNVISITED:
queue.append((y+dy, x+dx))
grid[y+dy][x+dx] = (y, x)
cur = goal
while cur != WALL:
y, x = cur
cur = grid[y][x]
grid[y][x] = PATH
def pt_2_ch(pt):
if pt == WALL: return '.'
elif pt == PATH: return '*'
else: return ' '
maze_solved = "\n".join(''.join(pt_2_ch(pt) for pt in row) for row in grid)
print(maze_solved)
# ############################## # # # # # ### # # ## # ## # ## # # # # # # # # # # # # # # # # # ## # # ## ## # # # # ### # # # # # # ## # # # # # # ## ## ## # # # # # # # ###################### ######### .*.............................. .** . ***. . ***. ...*****. . ..*. ..*.* ..*.* . .**. .***. . ***. *.*. . *.*. . .**. . ..*. . ..*..*.* . .*.* ...**.*******. . .**.*.*.. .*.*. .**.**. .. .. .. .***.******.*****.**. . . ......................*.........
初心者用：スタートからゴールへの経路を全て塗りつぶすアルゴリズムです。
#例題の迷路をリスト型で入力しておきます。
maze = """\
######## #
# #
# ###### #
# # # #
# #### # #
# # #
######## #
## # #
# # # #
# ########""".split("\n")
print('\nこれは入力された迷路です。')
for i in range(10): #入力した迷路を表示します。
print(maze[i])
print('\nこれは経路を全て+で塗りつぶしたものです。')
for i in range(10): #経路をすべて+で塗りつぶします。
maze[i] = maze[i].replace(' ','+')
print(maze[i])
def dig(mat,row,col):
st = mat[row-1][col] + mat[row][col+1] + mat[row+1][col] + mat[row][col-1]
cnt = st.count('#') + st.count(' ')
if cnt >= 3:
mat[row] = mat[row][:col] + ' ' + mat[row][col+1:]
else:
pass
def ent_dig():
for i in range(1,9):
for j in range(1,9):
if maze[i][j] == '+':
dig(maze,i,j)
cnt = 0
while True:
pre_maze = list(maze)
ent_dig()
if maze == pre_maze:
print('\n経路検出の結果です。')
for i in range(10):
print(maze[i])
break
else:
cnt += 1
print('\n{0}回目の誤経路削除の結果です。'.format(cnt))
for i in range(10):
print(maze[i])