将迷宫定义为数组,其中1是墙,0是可通行区域:
Must include start node in distance, if you BFS this it will give you 21.
[0][0] is the start point.
|
[ V
[0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 0],
[0, 0, 0, 0, 0, 0],
[0, 1, 1, 1, 1, 1],
[0, 1, 1, 1, 1, 1],
[0, 0, 0, 0, 0, 0]<-- [-1][-1] is the end point.
]
我们必须找到最短路径,我们可以删除一个'1'以帮助创建快捷方式。
创建最短路径的快捷方式是将[1] [0]更改为0,打开一条使距离为11的路径。
[
[0, 0, 0, 0, 0, 0],
-->[0, 1, 1, 1, 1, 0],
[0, 0, 0, 0, 0, 0],
[0, 1, 1, 1, 1, 1],
[0, 1, 1, 1, 1, 1],
[0, 0, 0, 0, 0, 0]
]
return 11
我的原始思维过程遍历每个元素并检查它是否= = 1,然后执行bfs将距离与最小值进行比较。
但是当然太慢了。所以我想要遍历每个元素并检查它是否为1,然后看看它是否有两个可通过的邻居,因为这似乎是唯一可能的情况,其中快捷方式是有意义的。
这是我的代码:
import copy
def bfs(maze):
visited = set()
queue = []
mazeHeight = len(maze)
mazeWidth = len(maze[0])
queue.append(((0,0),1))
while queue:
yx,distance = queue.pop(0)
y,x = yx
visited.add(yx)
if yx == (mazeHeight-1,mazeWidth-1):
return distance
if y+1 < mazeHeight:
if not maze[y+1][x] and (y+1,x) not in visited:
queue.append(((y+1,x),distance+1))
if y-1 >= 0:
if not maze[y-1][x] and (y-1,x) not in visited:
queue.append(((y-1,x),distance+1))
if x+1 < mazeWidth:
if not maze[y][x+1] and (y,x+1) not in visited:
queue.append(((y,x+1),distance+1))
if x-1 >= 0:
if not maze[y][x-1] and (y,x-1) not in visited:
queue.append(((y,x-1),distance+1))
return False
def answer(maze):
min = bfs(maze)
mazeHeight = len(maze)
mazeWidth = len(maze[0])
for y in range(mazeHeight):
for x in range(mazeWidth):
if maze[y][x]:
oneNeighbors = 0
if y+1 < mazeHeight:
if not maze[y+1][x]:
oneNeighbors += 1
if y-1 >= 0:
if not maze[y-1][x]:
oneNeighbors += 1
if x+1 < mazeWidth:
if not maze[y][x+1]:
oneNeighbors += 1
if x-1 >= 0:
if not maze[y][x-1]:
oneNeighbors += 1
if oneNeighbors == 2:
tmpMaze = copy.deepcopy(maze)
tmpMaze[y][x] = 0
tmpMin = bfs(tmpMaze)
if tmpMin < min:
min = tmpMin
return min
print(answer([[0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 0], [0, 0, 0, 0, 0, 0], [0, 1, 1, 1, 1, 1], [0, 1, 1, 1, 1, 1], [0, 0, 0, 0, 0, 0]]))
有什么建议可以提高速度吗?
答案 0 :(得分:1)
你似乎走在了正确的轨道上。可以考虑以下方法:
形成n x m
个节点的图表,其中n
和m
是迷宫矩阵的维度。
如果两个节点相邻0
s,则成本零的边缘。如果0
由1
分隔,则两个节点之间存在一的优势。
(请注意,每条路径需要维护两个成本,一个是上面的zero-one
成本,另一个是跟踪最小值的路径中的节点数。
然后执行BFS并仅考虑具有zero-one cost <= 1
的路径。
这将为您提供线性时间算法(节点数量为线性)。
以下代码可能包含错误,但它应该让您入门。
def bfs(maze):
visited = set()
queue = []
mazeHeight = len(maze)
mazeWidth = len(maze[0])
queue.append(((0,0),1,0))
while queue:
yx,distance, cost = queue.pop(0)
y,x = yx
visited.add(yx)
if yx == (mazeHeight-1,mazeWidth-1):
return distance
if y+1 < mazeHeight:
if not maze[y+1][x] and (y+1,x) not in visited:
queue.append(((y+1,x),distance+1, cost))
if y-1 >= 0:
if not maze[y-1][x] and (y-1,x) not in visited:
queue.append(((y-1,x),distance+1, cost))
if x+1 < mazeWidth:
if not maze[y][x+1] and (y,x+1) not in visited:
queue.append(((y,x+1),distance+1, cost))
if x-1 >= 0:
if not maze[y][x-1] and (y,x-1) not in visited:
queue.append(((y,x-1),distance+1, cost))
if cost == 0:
if y+2 < mazeHeight:
if not maze[y+2][x] and (y+2,x) not in visited and maze[y+1][x] == 1:
queue.append(((y+2,x),distance+2, cost+1))
if y-1 >= 0:
if not maze[y-2][x] and (y-2,x) not in visited and maze[y-1][x] == 1:
queue.append(((y-2,x),distance+2, cost+1))
if x+1 < mazeWidth:
if not maze[y][x+2] and (y,x+2) not in visited and maze[y][x+1] == 1:
queue.append(((y,x+2),distance+2, cost+1))
if x-1 >= 0:
if not maze[y][x-2] and (y,x-2) not in visited and maze[y][x-1] == 1:
queue.append(((y,x-2),distance+2, cost+1))
return False