Generate all connected graphs of order 12 with matching number of 4

asked 2024-04-20 16:49:49 +0200

licheng gravatar image

updated 2024-04-21 03:43:18 +0200

In fact, I can use the following code (in conjunction with Erdős and Gallai's theorem) where the number of edges is less than or equal to 38.

num_vertices = 12
selected_graphs = []
for g in graphs.nauty_geng(f'{num_vertices} -c 11:38'):
    if g.matching(value_only=True) == 4:
        selected_graphs.append(g)
len(selected_graphs)

I am wondering if it's possible to combine the structure of such graphs to improve efficiency. For example, Let $G$ be a connected graph with 12 vertices and maximum matchings with size of 4. Let $M=\{u_1u_2,u_3u_4,u_5u_6,u_7u_8\}$ be a matching. Let $S=V(G)-\{u_1,u_2\dots u_8\}$. Then:

  • there is no edge in $S$.
  • Any two vertices in $S$ cannot be adjacent to the distinct endpoints of a matching in M, respectively.

Perhaps in this way, we can need to traverse non-isomorphic graphs with smaller order, in this case only running some non-isomorphic graphs on 8 vertices. Or nauty can forbid matchings of fixed size, which could also achieve some purposes.

Doing this may be good for generating all such graphs with more vertices, such as 14 vertices. Actually, it might be more generally meaningful to generate non-isomorphic graphs containing a subgraph, and perhaps we can reduce the search space by using some basic structures.

PS:

Theorem (Erdős and Gallai, 1959) For all $n \geq 2 s$ and $s \geq 1$, $$ ex(n, M_s) = \max \{\binom{s - 1}{2} + (s - 1)(n - s + 1), \binom{2s - 1}{2} \}, $$ that is, for any given graph $G$ of order $ n$, if $|E(G)| > ex(n, M_s) $, then $ G $ contains a matching of size $s$.

edit retag flag offensive close merge delete

Comments

Why { cannot be shown in environment $$ on this platform.

licheng gravatar imagelicheng ( 2024-04-20 17:28:13 +0200 )edit
1

It can: $\{$ - use \\{

Max Alekseyev gravatar imageMax Alekseyev ( 2024-04-20 18:23:24 +0200 )edit

Did you get counts for graphs of order 10 or 11?

Max Alekseyev gravatar imageMax Alekseyev ( 2024-04-21 15:34:09 +0200 )edit

166988 for such connected graphs of order 10

licheng gravatar imagelicheng ( 2024-04-23 07:32:26 +0200 )edit

This suggests that it may be already challenging to get those graphs for order 11, while order 12 may be out of reach simply because the number of graphs of interest is prohibitively large.

Max Alekseyev gravatar imageMax Alekseyev ( 2024-04-23 12:59:36 +0200 )edit