Dynamic algorithm for graph connectedness?




Related to the recent thread on minimum vertex covers:
Does anyone here know a good algorithm or data structure for testing
whether a graph has been disconnected by the removal of an edge?

If we're looking for a minimum vertex cover, we're going to be
steadily removing edges from the graph. If the graph becomes
disconnected, then we can run the MVC algorithm independently on
each connected component, and that'll be faster than running a
brute-force MVC algorithm on the whole graph, because 2^x + 2^y
is less than 2^(x+y).

But to do that, we need to be notified when the graph becomes
disconnected. I can't figure out how to do that faster than O(n) ---
remove the edge (a,b), and then try to get from a to b using a
depth-first search. But doing an O(n) algorithm every time an edge
is removed sounds really inefficient!

If the graph is planar, and we already know how to embed it in the
plane, then there's a nice quick (O(lg n)?) algorithm: If we're removing
an edge, and that edge doesn't separate two different faces of the graph,
then we're disconnecting the graph.

What if the graph might not be planar, though?

This seems like a well-studied problem, but I can't find any algorithms
online.

-Arthur
.



Relevant Pages

  • Re: Minimum vertex cover algorithm in theory
    ... I am interested not in its performance, but in an algorithm that can ... Given a graph G =, how to find a minimum vertex cover for the ... Edge ending point count is decreased by 1; ...
    (comp.programming)
  • Re: Minimum vertex cover algorithm in theory
    ... propaul wrote: ... I am interested not in its performance, but in an algorithm that can ... Given a graph G =, how to find a minimum vertex cover for the ...
    (comp.programming)
  • Minimum vertex cover algorithm in theory
    ... I am interested not in its performance, but in an algorithm that can ... Given a graph G =, how to find a minimum vertex cover for the ... Edge ending point count is decreased by 1; ...
    (comp.programming)
  • Re: Hamberger Principle as a mind exercise
    ... it's quite easy to find graphs for which your algorithm does ... To remove any chance of success consider a seven vertex graph composed ... By "correctness," you mean that it really should find a minimum vertex ... That is why Mark says your algorithm is incorrect. ...
    (comp.programming)
  • Re: Shortish paths
    ... For how many nodes and how many edges will the actual algorithm run ... original graph. ... optimal solution is removed from the graph, ... I intend to begin by rereading the shortest path sections of Cormen ...
    (comp.theory)