You are viewing the master version of this library.
The master and develop pages are shown here to provide access to the newest in-development in-progress documentation. See the "Documentation" link on this page.
Other aspects of the page will reflect the previous official version, or not be available until the next release.
The Boost Graph Library (BGL) is a generic library that allows users to:
#include <boost/graph/adjacency_list.hpp>
#include <boost/graph/dijkstra_shortest_paths.hpp>
#include <boost/graph/visitors.hpp>
#include <iostream>
#include <limits>
#include <vector>
struct City {};
struct Road { int cost; };
using namespace boost;
using Graph = adjacency_list<vecS, vecS, directedS, City, Road>;
using Vertex = graph_traits<Graph>::vertex_descriptor;
int main() {
Graph g(4);
add_edge(0, 1, Road{1}, g);
add_edge(1, 2, Road{2}, g);
add_edge(0, 2, Road{10}, g);
add_edge(2, 3, Road{1}, g);
// Storage: you control allocation, lifetime, and container type
std::vector<Vertex> storage_pred(num_vertices(g));
std::vector<int> storage_dist(num_vertices(g));
// Property maps: lightweight views into the storage
auto index_map = get(vertex_index, g);
auto costs_map = get(&Road::cost, g);
auto predecessor_map = make_iterator_property_map(storage_pred.begin(), index_map);
auto distance_map = make_iterator_property_map(storage_dist.begin(), index_map);
dijkstra_shortest_paths(g, vertex(0, g),
predecessor_map, distance_map,
costs_map, index_map,
std::less<int>(), std::plus<int>(),
std::numeric_limits<int>::max(), 0,
dijkstra_visitor<null_visitor>());
for (auto v : make_iterator_range(vertices(g)))
std::cout << "distance to " << v << " = " << storage_dist[v] << "\n";
}
distance to 0 = 0
distance to 1 = 1
distance to 2 = 3
distance to 3 = 4
BGL ships dozens of graph algorithms: shortest paths (Dijkstra, Bellman-Ford, A*, Floyd-Warshall, Johnson), spanning trees (Kruskal, Prim), maximum flow (Edmonds-Karp, push-relabel, Boykov-Kolmogorov), traversal (BFS, DFS, topological sort), planarity testing, isomorphism, component decomposition, and more.
See the full algorithm reference for the complete catalogue.
[graph] tag in the subject line.#boost channel.Install Boost via your package manager:
| Manager | Command |
|---|---|
| vcpkg | vcpkg install boost-graph |
| Conan | conan install --requires=boost/[*] |
| apt (Debian/Ubuntu) | sudo apt install libboost-graph-dev |
| Homebrew (macOS) | brew install boost |
Then wire it into CMake:
find_package(Boost REQUIRED COMPONENTS graph)
target_link_libraries(my_app PRIVATE Boost::graph)
Most of BGL is header-only. Linking Boost::graph is only required for the GraphViz and GraphML parsers.
For working on BGL itself (building Boost from source, running the test suite), see CONTRIBUTING.md.
<h1>Boost Graph Library</h1> <p><a href="https://www.boost.org/doc/libs/release/libs/graph/doc/index.html"><img src="https://img.shields.io/badge/docs-boost.org-blue.svg" alt="Docs" /></a> <a href="https://en.cppreference.com/w/cpp/14"><img src="https://img.shields.io/badge/C%2B%2B-14-blue.svg" alt="C++14" /></a> <a href="https://github.com/boostorg/graph/actions/workflows/ci.yml"><img src="https://github.com/boostorg/graph/actions/workflows/ci.yml/badge.svg" alt="CI" /></a> <a href="https://www.boost.org/LICENSE_1_0.txt"><img src="https://img.shields.io/badge/License-BSL_1.0-blue.svg" alt="License: BSL-1.0" /></a> <a href="https://github.com/boostorg/boost/releases"><img src="https://img.shields.io/github/v/release/boostorg/boost?label=Boost&color=orange" alt="Boost release" /></a></p> <p>The Boost Graph Library (BGL) is a generic library that allows users to:</p> <ol> <li>Represent graph data using different structures (adjacency matrix, adjacency list, compressed sparse row, vectors of vectors, user-defined data structures).</li> <li>Attach user-defined data to vertices, edges, or the graph itself.</li> <li>Run a large number of algorithms on the graph.</li> <li>Inject user logic into algorithms using visitor hooks.</li> </ol> <h2>Example</h2> <p><a href="https://godbolt.org/z/9Esszr9Ga">Try it on Compiler Explorer</a></p> <pre class="highlightjs highlight"><code class="language-cpp hljs">#include <boost/graph/adjacency_list.hpp> #include <boost/graph/dijkstra_shortest_paths.hpp> #include <boost/graph/visitors.hpp> #include <iostream> #include <limits> #include <vector> struct City {}; struct Road { int cost; }; using namespace boost; using Graph = adjacency_list<vecS, vecS, directedS, City, Road>; using Vertex = graph_traits<Graph>::vertex_descriptor; int main() { Graph g(4); add_edge(0, 1, Road{1}, g); add_edge(1, 2, Road{2}, g); add_edge(0, 2, Road{10}, g); add_edge(2, 3, Road{1}, g); // Storage: you control allocation, lifetime, and container type std::vector<Vertex> storage_pred(num_vertices(g)); std::vector<int> storage_dist(num_vertices(g)); // Property maps: lightweight views into the storage auto index_map = get(vertex_index, g); auto costs_map = get(&Road::cost, g); auto predecessor_map = make_iterator_property_map(storage_pred.begin(), index_map); auto distance_map = make_iterator_property_map(storage_dist.begin(), index_map); dijkstra_shortest_paths(g, vertex(0, g), predecessor_map, distance_map, costs_map, index_map, std::less<int>(), std::plus<int>(), std::numeric_limits<int>::max(), 0, dijkstra_visitor<null_visitor>()); for (auto v : make_iterator_range(vertices(g))) std::cout << "distance to " << v << " = " << storage_dist[v] << "\n"; } </code></pre> <pre class="highlightjs highlight"><code class="language- hljs">distance to 0 = 0 distance to 1 = 1 distance to 2 = 3 distance to 3 = 4 </code></pre> <h2>Algorithms</h2> <p>BGL ships dozens of graph algorithms: shortest paths (Dijkstra, Bellman-Ford, A*, Floyd-Warshall, Johnson), spanning trees (Kruskal, Prim), maximum flow (Edmonds-Karp, push-relabel, Boykov-Kolmogorov), traversal (BFS, DFS, topological sort), planarity testing, isomorphism, component decomposition, and more.</p> <p>See the <a href="https://becheler.github.io/graph/graph/algorithms/overview.html">full algorithm reference</a> for the complete catalogue.</p> <h2>Help and feedback</h2> <ul> <li><strong><a href="https://github.com/boostorg/graph/issues">GitHub Issues</a></strong> for bug reports. Search before opening a new one.</li> <li><strong><a href="https://github.com/boostorg/graph/discussions">GitHub Discussions</a></strong> for questions, design ideas, and general conversation about the library.</li> <li><strong><a href="http://lists.boost.org/mailman/listinfo.cgi/boost-users">Boost mailing list</a></strong> for general Boost development. Use the <code>[graph]</code> tag in the subject line.</li> <li><strong>CppLang Slack</strong> for real-time chat. <a href="https://cppalliance.org/slack/">Request an invite</a>, then join the <code>#boost</code> channel.</li> <li><strong>Direct contact with maintainers</strong>: see <a href="CONTRIBUTING.md#maintainers">CONTRIBUTING.md#maintainers</a>.</li> </ul> <h2>Using BGL</h2> <p>Install Boost via your package manager:</p> <table> <thead> <tr> <th align="left">Manager</th> <th align="left">Command</th> </tr> </thead> <tbody> <tr> <td align="left"><a href="https://vcpkg.io">vcpkg</a></td> <td align="left"><code>vcpkg install boost-graph</code></td> </tr> <tr> <td align="left"><a href="https://conan.io">Conan</a></td> <td align="left"><code>conan install --requires=boost/[*]</code></td> </tr> <tr> <td align="left">apt (Debian/Ubuntu)</td> <td align="left"><code>sudo apt install libboost-graph-dev</code></td> </tr> <tr> <td align="left">Homebrew (macOS)</td> <td align="left"><code>brew install boost</code></td> </tr> </tbody> </table> <p>Then wire it into CMake:</p> <pre class="highlightjs highlight"><code class="language-cmake hljs">find_package(Boost REQUIRED COMPONENTS graph) target_link_libraries(my_app PRIVATE Boost::graph) </code></pre> <p>Most of BGL is header-only. Linking <code>Boost::graph</code> is only required for the GraphViz and GraphML parsers.</p> <h2>Building from source</h2> <p>For working on BGL itself (building Boost from source, running the test suite), see <a href="CONTRIBUTING.md">CONTRIBUTING.md</a>.</p>
<h1>Boost Graph Library</h1> <p><a href="https://www.boost.org/doc/libs/release/libs/graph/doc/index.html"><img src="https://img.shields.io/badge/docs-boost.org-blue.svg" alt="Docs" /></a> <a href="https://en.cppreference.com/w/cpp/14"><img src="https://img.shields.io/badge/C%2B%2B-14-blue.svg" alt="C++14" /></a> <a href="https://github.com/boostorg/graph/actions/workflows/ci.yml"><img src="https://github.com/boostorg/graph/actions/workflows/ci.yml/badge.svg" alt="CI" /></a> <a href="https://www.boost.org/LICENSE_1_0.txt"><img src="https://img.shields.io/badge/License-BSL_1.0-blue.svg" alt="License: BSL-1.0" /></a> <a href="https://github.com/boostorg/boost/releases"><img src="https://img.shields.io/github/v/release/boostorg/boost?label=Boost&color=orange" alt="Boost release" /></a></p> <p>The Boost Graph Library (BGL) is a generic library that allows users to:</p> <ol> <li>Represent graph data using different structures (adjacency matrix, adjacency list, compressed sparse row, vectors of vectors, user-defined data structures).</li> <li>Attach user-defined data to vertices, edges, or the graph itself.</li> <li>Run a large number of algorithms on the graph.</li> <li>Inject user logic into algorithms using visitor hooks.</li> </ol> <h2>Example</h2> <p><a href="https://godbolt.org/z/9Esszr9Ga">Try it on Compiler Explorer</a></p> <pre class="highlightjs highlight"><code class="language-cpp hljs">#include <boost/graph/adjacency_list.hpp> #include <boost/graph/dijkstra_shortest_paths.hpp> #include <boost/graph/visitors.hpp> #include <iostream> #include <limits> #include <vector> struct City {}; struct Road { int cost; }; using namespace boost; using Graph = adjacency_list<vecS, vecS, directedS, City, Road>; using Vertex = graph_traits<Graph>::vertex_descriptor; int main() { Graph g(4); add_edge(0, 1, Road{1}, g); add_edge(1, 2, Road{2}, g); add_edge(0, 2, Road{10}, g); add_edge(2, 3, Road{1}, g); // Storage: you control allocation, lifetime, and container type std::vector<Vertex> storage_pred(num_vertices(g)); std::vector<int> storage_dist(num_vertices(g)); // Property maps: lightweight views into the storage auto index_map = get(vertex_index, g); auto costs_map = get(&Road::cost, g); auto predecessor_map = make_iterator_property_map(storage_pred.begin(), index_map); auto distance_map = make_iterator_property_map(storage_dist.begin(), index_map); dijkstra_shortest_paths(g, vertex(0, g), predecessor_map, distance_map, costs_map, index_map, std::less<int>(), std::plus<int>(), std::numeric_limits<int>::max(), 0, dijkstra_visitor<null_visitor>()); for (auto v : make_iterator_range(vertices(g))) std::cout << "distance to " << v << " = " << storage_dist[v] << "\n"; } </code></pre> <pre class="highlightjs highlight"><code class="language- hljs">distance to 0 = 0 distance to 1 = 1 distance to 2 = 3 distance to 3 = 4 </code></pre> <h2>Algorithms</h2> <p>BGL ships dozens of graph algorithms: shortest paths (Dijkstra, Bellman-Ford, A*, Floyd-Warshall, Johnson), spanning trees (Kruskal, Prim), maximum flow (Edmonds-Karp, push-relabel, Boykov-Kolmogorov), traversal (BFS, DFS, topological sort), planarity testing, isomorphism, component decomposition, and more.</p> <p>See the <a href="https://becheler.github.io/graph/graph/algorithms/overview.html">full algorithm reference</a> for the complete catalogue.</p> <h2>Help and feedback</h2> <ul> <li><strong><a href="https://github.com/boostorg/graph/issues">GitHub Issues</a></strong> for bug reports. Search before opening a new one.</li> <li><strong><a href="https://github.com/boostorg/graph/discussions">GitHub Discussions</a></strong> for questions, design ideas, and general conversation about the library.</li> <li><strong><a href="http://lists.boost.org/mailman/listinfo.cgi/boost-users">Boost mailing list</a></strong> for general Boost development. Use the <code>[graph]</code> tag in the subject line.</li> <li><strong>CppLang Slack</strong> for real-time chat. <a href="https://cppalliance.org/slack/">Request an invite</a>, then join the <code>#boost</code> channel.</li> <li><strong>Direct contact with maintainers</strong>: see <a href="CONTRIBUTING.md#maintainers">CONTRIBUTING.md#maintainers</a>.</li> </ul> <h2>Using BGL</h2> <p>Install Boost via your package manager:</p> <table> <thead> <tr> <th align="left">Manager</th> <th align="left">Command</th> </tr> </thead> <tbody> <tr> <td align="left"><a href="https://vcpkg.io">vcpkg</a></td> <td align="left"><code>vcpkg install boost-graph</code></td> </tr> <tr> <td align="left"><a href="https://conan.io">Conan</a></td> <td align="left"><code>conan install --requires=boost/[*]</code></td> </tr> <tr> <td align="left">apt (Debian/Ubuntu)</td> <td align="left"><code>sudo apt install libboost-graph-dev</code></td> </tr> <tr> <td align="left">Homebrew (macOS)</td> <td align="left"><code>brew install boost</code></td> </tr> </tbody> </table> <p>Then wire it into CMake:</p> <pre class="highlightjs highlight"><code class="language-cmake hljs">find_package(Boost REQUIRED COMPONENTS graph) target_link_libraries(my_app PRIVATE Boost::graph) </code></pre> <p>Most of BGL is header-only. Linking <code>Boost::graph</code> is only required for the GraphViz and GraphML parsers.</p> <h2>Building from source</h2> <p>For working on BGL itself (building Boost from source, running the test suite), see <a href="CONTRIBUTING.md">CONTRIBUTING.md</a>.</p>