|
||
Title: Counting Graphs Post by vphan on Jul 6th, 2004, 12:01pm hello all, Any ideas to the following problem? how many acyclic directed graphs with n vertices are there? how many of them are also transitive closures? Any ideas? |
||
Title: Re: Counting Graphs Post by Noke Lieu on Jul 6th, 2004, 7:21pm How's the course going? There was just something about the way you asked that this seems like homework. Sadly, I am of no help here whatsoever. I'd trawl through mathworld (http://mathworld.wolfram.com/AcyclicDigraph.html) until I felt I understood what the question was asking, and by then one of the seriously clever inhabitants here would come and do a far better job than I ever could. Hold tight, the answer will come. [hide]But probably in a while to help miss your deadline.[/hide] |
||
Title: Re: Counting Graphs Post by william wu on Jul 27th, 2004, 5:49am I have to return to my homework soon, but for the first question, here's a possible approach ... Note that we can represent any graph as a 0-1 adjacency matrix A, where entry (i,j) is 1 if there is a directed edge from i to j, and 0 otherwise. (For undirected graphs, the adjacency matrix is symmetric.) Then here are some facts: 1) Entry (i,j) in Ak corresponds to the number of paths of length k from node i to node j. 2) (I - A)-1 = I + A + A2 + ... Combining 1) and 2), we see that (I-A) is invertible iff there are no cycles in the graph. (Recognize that if there is a cycle containing nodes i and j, then there are paths of every possible length from i to j.) Then, to find the number of acyclic directed graphs with n vertices, one could count the number of distinct 0-1 matrices "A" such that (I - A) is invertible. Not sure if this is feasible, but seems easier to think about for me than the original problem. Also, this assumes you are counting labeled graphs; otherwise you will have to count classes of similar matrices (isomorphic graphs). |
||
Powered by YaBB 1 Gold - SP 1.4! Forum software copyright © 2000-2004 Yet another Bulletin Board |