[英]Create undirected graph by boost graph same as the input
我想按如下方式创建图形,第一列是顶点,其他是邻接顶点
我这样将边添加到图形中
using MinCutG = boost::adjacency_list<boost::vecS, boost::vecS, boost::undirectedS>;
MinCutG graph;
std::vector<std::vector<int> > input{ {1,2,3,4,7}, {2,1,3,4}, {3,1,2,4},
{4,1,2,3,5}, {5,4,6,7,8}, {6,5,7,8}, {7,1,5,6,8}, {8,5,6,7}};
for(auto const &data : input){
auto begin = std::begin(data);
auto end = std::end(data);
if(begin != end){
auto const Vertex = *begin;
++begin;
while(begin != end){
boost::add_edge(Vertex, *begin, graph);
++begin;
}
}
}
打印图
template<typename Graph>
void print_adjacent_vertex(Graph const &g)
{
for (auto vertex = vertices(g); vertex.first != vertex.second; ++vertex.first){
std::cout << *vertex.first << " is connected with ";
for (auto neighbour = adjacent_vertices(*vertex.first, g);
neighbour.first != neighbour.second; ++neighbour.first){
std::cout << *neighbour.first << " ";
}
std::cout<<std::endl;
}
}
我期望输出应与输入相同,但不是结果是
我的预期结果
简而言之,将setS
用作OutEdgeList
模板参数将禁用平行边,从而得到所需的输出。
boost::adjacency_list
的第一个模板参数是OutEdgeList
,它控制某些图形行为,例如允许或不允许平行边。 在无向MinCutG
图形的情况下,将vecS
用作OutEdgeList
,它将启用平行边。 例如,如果无向图支持平行边,则使用:
add_edge(1, 2, graph); // Vertex 1 and 2 are adjacent to one another via edge A.
add_edge(2, 1, graph); // Vertex 1 and 2 are adjacent to one another via edge B,
// which is parallel to edge A.
顶点1
的adjacent_vertices()
将包含两次顶点2
,每个边( A
和B
)一次。
如文档中所述,可以通过使用OutEdgeList
setS
或hash_setS
选择器来禁用平行边。 例如,更改:
using MinCutG = boost::adjacency_list<
boost::vecS, // OutEdgeList with parallel edges
boost::vecS,
boost::undirectedS>;
至:
using MinCutG = boost::adjacency_list<
boost::setS, // OutEdgeList with no parallel edges
boost::vecS,
boost::undirectedS>;
下面是一个例子使用原始代码和只改变OutEdgeList
从vecS
到setS
:
#include <iostream>
#include <vector>
#include <boost/graph/adjacency_list.hpp>
template<typename Graph>
void print_adjacent_vertex(Graph const &g)
{
for (auto vertex = vertices(g); vertex.first != vertex.second;
++vertex.first){
std::cout << *vertex.first << " is connected with ";
for (auto neighbour = adjacent_vertices(*vertex.first, g);
neighbour.first != neighbour.second; ++neighbour.first){
std::cout << *neighbour.first << " ";
}
std::cout<<std::endl;
}
}
int main()
{
using MinCutG = boost::adjacency_list<
boost::setS, boost::vecS, boost::undirectedS>;
MinCutG graph;
std::vector<std::vector<int> > input
{
{1,2,3,4,7},
{2,1,3,4},
{3,1,2,4},
{4,1,2,3,5},
{5,4,6,7,8},
{6,5,7,8},
{7,1,5,6,8},
{8,5,6,7}
};
for(auto const &data : input){
auto begin = std::begin(data);
auto end = std::end(data);
if(begin != end){
auto const Vertex = *begin;
++begin;
while(begin != end){
boost::add_edge(Vertex, *begin, graph);
++begin;
}
}
}
print_adjacent_vertex(graph);
}
产生以下输出:
0 is connected with
1 is connected with 2 3 4 7
2 is connected with 1 3 4
3 is connected with 1 2 4
4 is connected with 1 2 3 5
5 is connected with 4 6 7 8
6 is connected with 5 7 8
7 is connected with 1 5 6 8
8 is connected with 5 6 7
正如Tanner所解释的,实际上每个边缘都加了两次,首先是(a,b),然后是(b,a)。 避免这种情况的一种但相当昂贵的方法是使用setS
作为控制边缘容器的模板参数。 这就是Tanner在他的回答中所建议的。
另一种方法是在添加边缘之前检查边缘是否存在。 这意味着线
boost::add_edge(Vertex, *begin, graph);
应该替换为
if (! edge(Vertex, *begin, graph).second ) //edge does not exist yet
boost::add_edge(Vertex, *begin, graph);
在大多数情况下,此代码可能会更快并且消耗更少的内存。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.