Analysis of Connectivity in Graphs - PowerPoint PPT Presentation

1 / 10
About This Presentation
Title:

Analysis of Connectivity in Graphs

Description:

Obtain a histogram of vertex-connectivity for all graphs of order = 10. Compute minimum number of edges needed to have a ... Similar to Dinic's algorithm ... – PowerPoint PPT presentation

Number of Views:64
Avg rating:3.0/5.0
Slides: 11
Provided by: wen153
Category:

less

Transcript and Presenter's Notes

Title: Analysis of Connectivity in Graphs


1
Analysis of Connectivity in Graphs
  • Harini Krishnamurthy
  • Karthik Nandakumar
  • Kuo-Liang Chang

2
Goals
  • Obtain a histogram of vertex-connectivity for all
    graphs of order lt 10
  • Compute minimum number of edges needed to have a
    k-connected n-order graph
  • Analyze the percentage of connected graphs that
    satisfy the equality in Whitneys theorem
  • i.e. ? ? ?

3
Finding Connectivity of a graph
Initialize ? n Form the network N G (V, E,
u, v).
Get adjacency matrix from nauty
Find Max-Flow in N, f(N)
if f(N) lt ? ? f(N)
4
Transformation of Graphfrom graph to the
corresponding network

x
x-
y-
s-
s
t
t-
Source
Sink
y
5
Max-Flow Algorithm
  • An Experimental Comparison of Min-cut / Max-flow
    Algorithms for Energy Minimization in Vision
  • - Yuri Boykov and Vladimir Kolmogorov
  • Similar to Dinics algorithm
  • Builds search trees to find the augmenting paths
    but reuses this tree for further iterations
  • Drawback does not find the shortest augmenting
    path
  • Theoretically, time complexity of this algorithm
    is O(mn2) but authors have experimentally shown
    that on typical problem instances in vision, it
    is much faster compared to other standard
    algorithms.

6
From Max-Flow to Vertex Connectivity
  • If G is a complete graph, ? n-1 and stop
  • If G is disconnected, ? 0 and stop
  • ? n
  • i 0
  • Step to determine if connectivity has been
    computed
  • if i ? ?, then i i1, continue. Else
    stop.
  • 6. j i1
  • 7. Update ? until ?(G) is found
  • 7.1 if j n1, return to step 5
  • 7.2 if uv E(G), construct network N
    (V, E, u, v-)
  • 7.3 Determine f(N).
  • 7.4 if f(N) lt ?, ? f(N), else
    continue
  • 7.5 j j1. Return to step 7.1

7
Histogram of connectivity (1)
disconnected
1 connected
2 connected
3 connected
4 connected
8
Histogram of connectivity (2)
9
Equality in Whitneys theorem? ? ?
10
Observations
  • As the order n increases, the percentage of
    graphs having lower connectivity decreases.
  • The minimum number of edges needed to have a
    k-connected n-order graph is
  • n ?(k-2)n/2?
  • (If k 1, then it is n-1)
  • Of all connected graphs of order, n lt 10,
  • nearly 96 satisfy the equality ? ? ? and
  • nearly 97.5 of the graphs have ? ?
Write a Comment
User Comments (0)
About PowerShow.com