如何在Networkx中指定边长以计算最短距离? [英] How to specify edge length in Networkx for calculating shortest distance?

问题描述

我有节点和边的列表，但我希望某些边的长度为2，而不是1.因此，当使用内置算法计算节点之间的距离时，它将返回

I have a list of nodes and edges but I want some edges to be of length two instead of one. So when the distance between the nodes is calculated using the built in algorithm it returns

例如，如果我以(1、2)，(2 *，3)，(4 *，5)作为节点之间的边，其中带有星号的节点之间的距离为长度2，则( 1、2)应该是1，(2,3)应该是2而不是1，然后(1,5)之间的距离应该是5而不是3.

For example, if I have (1, 2), (2*, 3), (4*, 5) as edges between nodes, where the distance between the nodes with an asterisk have length two, the distance between (1, 2) should be 1, (2,3) should be 2 instead of 1 and then the distance between (1,5) should be 5 instead of 3.

添加节点时，我尝试了G.add_edge(4,5,length=2)，但nx.shortest_path_length(G,source=4,target=5))仍然返回1而不是2.如何指定边缘长度?

When adding nodes I've tried G.add_edge(4,5,length=2) but nx.shortest_path_length(G,source=4,target=5)) still returns 1 instead of two. How can I specify edge length?

推荐答案

您需要在边缘上附加一个length属性，然后指定查找最短路径时要按这些长度进行加权:

You need to have a length attribute attached to your edges, and then specify that you want to weight by those lengths when finding the shortest path:

# Had to add an edge from 3 to 4 to your example edges
#   or there's no path from 1 to 5
edges = [(1, 2, 1), (2, 3, 2), (3, 4, 1), (4, 5, 2)]

G = networkx.Graph()

for start, end, length in edges:
    # You can attach any attributes you want when adding the edge
    G.add_edge(start, end, length=length)

networkx.shortest_path_length(G, 1, 5, weight='length')
Out[8]: 6

这篇关于如何在Networkx中指定边长以计算最短距离?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！