Graphs and digraphs with modern C++

General

• Code for weighted and unweighted graph and digraph problems
• No usage of owning raw pointers, new etc. (uses std::unique_ptr, std::vector, ... instead)
• Unit tests in test/ subdirectories
• This is a sample project and not a huge library. It is currently not split in headers and implementations.

General headers (general/)

include/

• DisjointSets as an efficient union-find data structure with path compression and union-by-rank (see also here)
• IndexedPriorityQueue as a priority queue which supports changing keys in logarithmic time (see here), and PriorityQueue as a simpler version which does not

Unweighted graphs (unweighted_graph/)

include/

• Simple (undirected) Graph interface
• AdjacencyListGraph (https://en.wikipedia.org/wiki/Adjacency_list)
• Basic utility functions (a few)
• Find paths starting at a vortex to all connected vertices
  • with Depth-first Search (DFS) recursively (graph::find_paths_to_all::fromVertexToAllDfs)
  • with DFS iteratively (graph::find_paths_to_all::fromVertexToAllDfsNoRec)
  • with Breadth-first Search (BFS) iteratively (graph::find_paths_to_all::fromVertexToAllBfs)
• Find connected components (https://en.wikipedia.org/wiki/Component_(graph_theory)) with DFS (graph::ConnectedComponents)
• Check whether a graph is bipartite with DFS (graph::isBipartite)

unweighted_graph_demo.cpp

• Basic test of Graph.h functionality
• Reads graphs from standard input (build/unweighted_graph/unweighted_graph_demo < unweighted_graph/tinyG.txt) or file given as program argument (build/unweighted_graph/unweighted_graph_demo unweighted_graph/tinyG.txt)
• Sample graph from tinyG.txt:

Unweighted digraphs (unweighted_digraph/)

include/

• Simple Digraph interface
• AdjacencyListDigraph (https://en.wikipedia.org/wiki/Adjacency_list)
• check for cycles
• determine topological sort (https://en.wikipedia.org/wiki/Topological_sorting)
• determine strongly-connected components with Kosaraju-Sharir algorithm (https://en.wikipedia.org/wiki/Kosaraju%27s_algorithm)

unweighted_digraph_demo.cpp

• Basic test of Digraph.h functionality

• Reads digraph from standard input (build/unweighted_digraph/unweighted_digraph_demo < unweighted_digraph/tinyDG.txt) or file given as program argument (build/unweighted_digraph/unweighted_digraph_demo unweighted_digraph/tinyDG.txt)

• Sample digraph from tinyDG.txt:

  

Weighted graphs (weighted_graph/)

include/

• EdgeWeightedGraph interface and an implementation with adjacency lists
• Minimum Spanning Tree identification with Kruskal's algorithm

weighted_graph_demo.cpp

• Reads a weighted graph from standard input or a file given as program argument

• Prints the graph's edges and the Minimum Spanning Tree edges

• Sample edge-weighted graph from tinyEWG.txt and its Minimum Spanning Tree:

  

Weighted digraphs (weighted_digraph/)

include/

• EdgeWeightedDigraph interface and an implementation with adjacency lists
• SingleSourceShortestPath as an interface for finding shortest paths to all other nodes starting at a start vertex:
  • SingleSourceAcyclicShortestPath for acyclic weighted digraphs using the topological order
  • SingleSourceDijkstraShortestPath as an implementation of Dijkstra's algorithm for weighted digraphs without negative edge weights
  • SingleSourceBellmanFordShortestPath for digraphs with negative edge weights (but without negative cycles)

weighted_digraph_demo.cpp

• Reads a weighted digraph from standard input or a file given as program argument

• Determines the shortest paths starting at vertex 0 and prints the results

  • Depending on whether the digraph is acyclic, contains no negative edge weights, or does contain negative edge weights, the DAG algorithm, Dijkstra's algorithm, or the Bellman-Ford algorithm is chosen respectively

• Sample edge-weighted digraph from tinyEWD.txt (without showing weights):

  

• Sample edge-weighted acyclic digraph from tinyEWDAG.txt (without showing weights):

  

Flow networks (flow_network/)

include/

• FlowEdge as an edge with capacity and flow in a flow network
• FlowNetwork and AdjancyListFlowNetwork as interface and implementation of a flow network
• FordFulkerson as an implementation of the Ford-Fulkerson algorithm to solve the min-cut and max-flow problems in flow networks

flow_network_demo.cpp

• Reads flow network from file or standard input and prints it

• Calculate max-flow and min-cut

  

Compilation and execution

• Download submodules (for unit tests): git submodule update --init --recursive
• Compile with cmake:

  mkdir build
  cd build/
  cmake ..
  make
• Go to top-level folder again: cd ..
• Run all tests: find build/ -name "*_gtest" -exec {} \;
• Run demos:
  • build/unweighted_graph/unweighted_graph_demo unweighted_graph/tinyG.txt
  • build/unweighted_digraph/unweighted_digraph_demo unweighted_digraph/tinyDG.txt
  • build/weighted_graph/weighted_graph_demo weighted_graph/tinyEWG.txt
  • build/weighted_digraph/weighted_digraph_demo weighted_digraph/tinyEWD.txt
  • build/weighted_digraph/weighted_digraph_demo weighted_digraph/tinyEWDAG.txt
  • build/flow_network/flow_network_demo flow_network/tinyFN.txt

References

• Introduction to Algorithms by Cormen et al.
• Algorithms, Part II by Princeton University (Robert Sedgewick, Kevin Wayne)