jorants / MV-matching-C

C implementation of the MV-matching algorithm

REDIRECT: https://github.com/jorants/MV-Matching-V2

This git repository used to contain an implementation of the Micali Vazirani Maximum Cardinality Matching Algorithm.

The code is currently in an early stage and although the algorithm is implemented and seems to work, it really isn't user friendly.

It may currently be used as a C library, and it is possible to make pictures of the algorithm as it progresses. There are currently no bugs or show-stoppers that we know of, though the library could still use several improvements.

A faster, optimized implementation of the algorithm appears in https://github.com/jorants/MV-Matching-V2

Obtaining the code and compiling it

To get the latest version of the code, you go to a terminal and do:

$ git clone https://github.com/jorants/MV-matching-C.git

And then:

$ cd MV-matching-C
$ make

Typical usage of libmv

Let's call G to the graph you are interested in knowing the maximum cardinality matching of. Begin by including the libmv.h header:

#include "libmv.h"

Now declare some unsigned int variable:

unsigned int size;

And do some stuff until size contains the number of vertices in G

Then initialize a graph structure:

Graph *graph = Graph_init(size)

Now, make sure you assign a number from 0 to size-1 to each vertex in G. Maybe you already have this, but maybe your graph vertices are named some other way. In any case, enumerate them from 0 to size-1.

Then do some loop or something over your graph data, and for each edge (v1, v2) in G, call the function:

   Graph_add_edge(graph, v1, v2);

(Of course, v1, v2 are the numbers identifying the vertices in the edge.)

When you've done that for all edges, call:

EdgeList *lst = MV_MaximumCardinalityMatching(graph);

At this point, lst is a list of edges containing a maximum cardinality matching. To go through the individual edges in this list, use an iterator, as follows:

EdgeListIterator *itr = lst->first;
while (itr != 0) {
	Edge edge = itr->value;
	
	... do stuff with edge, maybe store it in your own data structure or whatever ...
	
	itr = itr->next;
}

To clear up the memory that was allocated, call:

EdgeList_delete(lst);
Graph_delete(graph);

And that's it.