08.10.2016 Views

Foundations of Data Science

2dLYwbK

2dLYwbK

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

i<br />

↑<br />

↓<br />

↑<br />

↓<br />

↑<br />

j<br />

↑<br />

=⇒<br />

i<br />

⇐=<br />

↑<br />

↓<br />

↑<br />

↓<br />

↑<br />

j<br />

↑<br />

Insert current at each vertex<br />

equal to degree <strong>of</strong> the vertex.<br />

Extract 2m at vertex j. v ij = h ij<br />

(a)<br />

↑ ↑<br />

i<br />

j<br />

=⇒<br />

↓<br />

↓<br />

↓ ↓<br />

Reverse currents in (b).<br />

For new voltages −v ji = h ji.<br />

Since −v ji = v ij , h ji = v ij .<br />

(c)<br />

Extract current from i instead <strong>of</strong> j.<br />

For new voltages v ji = h ji .<br />

(b)<br />

2m i<br />

j<br />

2m<br />

=⇒ =⇒<br />

Superpose currents in (a) and (c).<br />

2mr ij = v ij = h ij + h ji = commute(i, j)<br />

(d)<br />

Figure 5.8: Illustration <strong>of</strong> pro<strong>of</strong> that commute(x, y) = 2mr xy where m is the number <strong>of</strong><br />

edges in the undirected graph and r xy is the effective resistance between x and y.<br />

Corollary 5.11 For vertices x and y in an n vertex graph, the commute time, commute(x, y),<br />

is less than or equal to n 3 .<br />

Pro<strong>of</strong>: By Theorem 5.9 the commute time is given by the formula commute(x, y) =<br />

2mr xy where m is the number <strong>of</strong> edges. In an n vertex graph there exists a path from x<br />

to y <strong>of</strong> length at most n. Since the resistance can not be greater than that <strong>of</strong> any path<br />

from x to y, r xy ≤ n. Since the number <strong>of</strong> edges is at most ( n<br />

2<br />

( n<br />

commute(x, y) = 2mr xy ≤ 2 n<br />

2)<br />

∼ = n 3 .<br />

)<br />

Again adding edges to a graph may increase or decrease the commute time. To see<br />

this consider three graphs: the graph consisting <strong>of</strong> a chain <strong>of</strong> n vertices, the graph <strong>of</strong><br />

Figure 5.7, and the clique on n vertices.<br />

Cover time<br />

The cover time, cover(x, G) , is the expected time <strong>of</strong> a random walk starting at vertex x<br />

in the graph G to reach each vertex at least once. We write cover(x) when G is understood.<br />

169

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!