FKT algorithm and adjacency matrices

447 Views Asked by At

The Wikipedia article on the FKT algorithm says that one finds the number of perfect matchings in an undirected planar graph $G$ as follows.

  1. Find a graph $H$ that is a directed version of $G$, such that an odd number of edges are oriented clockwise for every face in $G$.

  2. Calculate $B$, the $(1, −1, 0)$-adjacency matrix of $H$.

  3. The perfect number of matchings for $G$ is $\rm{pf}(B) = \sqrt{\det(B)}$.

The hard part is step 1, which is what FKT deals with.

I have two questions:

  1. Is my above summary correct?

  2. What does it mean by the $(1, −1, 0)$-adjacency matrix? From the article it links to, I would guess that you put $1$ for each edge, $-1$ for the lack of an edge and $0$ on the diagonal. But this would not yield a skew symmetric $B$. Is it instead $1$ along the direction of an edge, $-1$ against, and $0$ for none?

1

There are 1 best solutions below

0
On BEST ANSWER

Assume that the vertex set of $G$ is $\{1,\ldots,n\}$. If $i$ and $j$ are not adjacent, set $B_{i,j}=0$. If $i$ and $j$ are adjacent and $i<j$, set $B_{i,j}=+1$ and $B_{j,i}=-1$; if $i$ and $j$ are adjacent and $i>j$ set $B_{i,j}=-1$ and $B_{j,i}=+1$. Now $B$ is skew symmetric and $\det(B)$ is the square of the number of perfect matchings of $G$.