try using dynamic programming approach

k length path from v = k-1 length subproblem from all of v's neighbours.

(and yes you will have to keep some pointers to ensure that it remains a
path (if u want simple path).


-Rohit


On Sun, Dec 27, 2009 at 8:35 AM, me13013 <[email protected]> wrote:

> The best solution is going to depend on properties of the graph.  One
> solution not yet mentioned is to perform breadth first searches from
> both v and N (if it's a digraph, backward search from N) and report
> paths when some node appears in the current "wave" from both ends.  If
> |V| >> the number of nodes we'd expect to visit, this will keep the
> number of nodes visited much lower than (roughly 2sqrt of) doing depth
> first search from v, enough to offset the additional memory
> requirements.
>
> Bob H
>
> On Dec 17, 6:49 am, vicky <[email protected]> wrote:
> > given a graph G(V,E) and a source vertex, v. no. of vetices n, and
> > edges e. you have to find  all different paths from vertex v to some
> > vertex N, having exactly i(1<i<n) edges. v,N,i will be given by the
> > user. provide an algorithm for it.
>
> --
>
> You received this message because you are subscribed to the Google Groups
> "Algorithm Geeks" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected]<algogeeks%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/algogeeks?hl=en.
>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Algorithm Geeks" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/algogeeks?hl=en.

Reply via email to