嘿喜欢标题说我正在尝试在Python3中实现深度限制搜索,它返回给定图形,起始顶点和目标顶点的路径。我正在努力解决如何强制执行搜索限制的问题。到目前为止,我有:
def dfs(g, v, goal, limit=-1):
SENTINEL = object()
visitedStack = [v]
path = ""
while visitedStack:
currentVertex = visitedStack.pop()
if g.getVertex(currentVertex) != None:
if g.getVertex(currentVertex).visited == False:
path += currentVertex + ' -> '
g.getVertex(currentVertex).hasBeenVisited()
if currentVertex == goal:
return path[:-3]
elif currentVertex == SENTINEL:
limit += 1
elif limit != 0:
limit -= 1
visitedStack.append(SENTINEL)
visitedStack.extend(g.getVertex(currentVertex).getConnections())
return "Depth limit was reached"
编辑:我更改了一些代码以检查访问过的顶点。编辑后,返回的搜索有时无法正常工作。例如,我将深度限制设置为3,但返回的路径为4或5。其他时候我会将限制设置为7并返回“达到限制”。注意:最小路径为3
答案 0 :(得分:1)
当搜索进入更深层次时,将哨兵推入堆栈并减少限制。当您从堆栈中弹出一个标记时,会增加该级别。
def dfs_limit(g, start, goal, limit=-1):
'''
Perform depth first search of graph g.
if limit >= 0, that is the maximum depth of the search.
'''
SENTINEL = object()
visitedStack = [start]
path = []
while visitedStack:
currentVertex = visitedStack.pop()
if currentVertex == goal:
path.append(currentVertex)
return ' -> '.join(path)
elif currentVertex == SENTINEL:
#finished this level; go back up one level
limit += 1
path.pop()
elif limit != 0:
# go one level deeper, push sentinel
limit -= 1
path.append(currentVertex)
visitedStack.append(SENTINEL)
visitedStack.extend(g.getVertex(currentVertex).getConnections())
如果图表中有循环或多条路线,您还需要跟踪已访问过的节点,这样您就不会重复工作或进入无限循环。