当图是有向的时候要使用dfs,这些是图的特性,应该在一开始有图做题的时候就应该有所判断:
下面是1559. 二维网格图中探测环 无向图dfs的范例:
sys.setrecursionlimit(999999999) class Solution: def containsCycle(self, grid: List[List[str]]) -> bool: direction = [(-1,0),(1,0),(0,-1),(0,1)] N,M = len(grid),len(grid[0]) explored=set() def dfs(node,pre): if node in explored: return True explored.add(node) for d in direction: x,y = node[0]+d[0],node[1]+d[1] if x<0 or x>=N or y<0 or y>=M or grid[x][y]!=grid[node[0]][node[1]] or (x,y)==pre: continue if dfs((x,y),node): return True return False for i in range(N): for j in range(M): if (i,j) not in explored and dfs((i,j),None): return True return False
一般情况下由于dfs 的时间复杂度为(b^d),效率会比较低,一般会加上剪枝的思考结果会比较理想。
解释
那么有向图又是怎么做的呢?
其实这种情况下一般我更推荐拓扑排序,但是实际使用起来,笔试的时候dfs会更容易想到。
经典的题目是leetcode 207课程表:
1是拓扑排序:即贪婪加bfs的解法:
class Solution: def canFinish(self, numCourses: int, prerequisites: List[List[int]]) -> bool: #课程的长度 clen = len(prerequisites) if clen ==0: return True in_degrees = [0 for _ in range(numCourses)] adj = [set() for _ in range(numCourses)] for second , first in prerequisites: in_degrees[second]+=1 adj[first].add(second) #print('in_degrees',in_degrees) #首先先遍历一遍 res=[] queue=[] for i in range(numCourses): if in_degrees[i]==0: queue.append(i) counter =0 #queue里面存的都是已经没有前置课的文章 while queue: top= queue.pop(0) counter+=1 for successor in adj[top]: in_degrees[successor]-=1 if in_degrees[successor] == 0: queue.append(successor) return counter==numCourses
dfs容易超时,这里建议使用flag标记已经遍历过的点或者用cache存一下结果:
class Solution: def canFinish(self, numCourses: int, prerequisites: List[List[int]]) -> bool: # 0 表示没有访问过(白) # 1 表示访问过了(黑) visited = [0] * numCourses adjacency = [[] for _ in range(numCourses)] @functools.lru_cache(None) def dfs(i): if visited[i] == 1: return False # if visited[i]==-1: # return True visited[i] = 1 for j in adjacency[i]: if not dfs(j): return False visited[i] = 0 return True for cur, pre in prerequisites: adjacency[cur].append(pre) for i in range(numCourses): if not dfs(i): return False return True
大部分问题下dfs会比较直观