Graphs  

Part  3:  applications  of  DFS  

Application  1  DFS:  topological  sort:    
A  directed  edge  from  u  to  v,  means  that  v  happens  after  u.  Topological  order:  All  
directed  edges  only  go  forward  (needs  to  be  acyclic  graph).  When  we  have  tasks  
that  one  needs  to  be  done  before  another.  Again  linear  time  in  vertices  and  
edges  O(n+m),  since  doing  DFS.  
22.4 Topological sort 613

11/16 undershorts socks 17/18

watch 9/10
12/15 pants shoes 13/14
shirt 1/8
(a) 6/7 belt
tie 2/5

jacket 3/4

(b) socks undershorts pants shoes watch shirt belt tie jacket
17/18 11/16 12/15 13/14 9/10 1/8 6/7 2/5 3/4

Figure 22.7 (a) Professor Bumstead topologically sorts his clothing when getting dressed. Each
directed edge .u; !/ means that garment u must be put on before garment !. The discovery and
finishing times from a depth-first search are shown next to each vertex. (b) The same graph shown
topologically sorted, with its vertices arranged from left to right in order of decreasing finishing time.
All directed edges go from left to right.

pants). A directed edge .u; !/ in the dag of Figure 22.7(a) indicates that garment u  
must be donned before garment !. A topological sort of this dag therefore gives an
Pseudo  code:  Topological  Sort  (G)  
order for getting dressed. Figure 22.7(b) shows the topologically sorted dag as an
ordering of vertices along a horizontal line such that all directed edges go from left
1.  Call  DFS(G)  to  compute  finish  times  v.f  for  each  vertex  v  
to right.
The following simple algorithm topologically sorts a dag:
2.  As  each  vertex  is  finished,  insert  into  the  front  of  a  linked  list  
3.  return  the  linked  list  of  vertices  
T OPOLOGICAL -S ORT .G/
Application  2  DFS:  Strongly  Connected  Components  
1 call DFS.G/ to compute finishing times !:f for each vertex !
2 as each vertex is finished, insert it onto the front of a linked list
This  is  another  example  application  of  Depth  First  Search  (DFS).  
3 return the linked list of vertices

Figure 22.7(b) shows how the topologically sorted vertices appear in reverse order
Definition:  Maximal  set  of  vertices  for  a  directed  graph,  so  that  for  each  pair  of  
of their finishing times.
vertices,  u  and  v,  there  is  a  path  both  from  u  to  v  and  from  v  to  u.  
We can perform a topological sort in time ‚.V C E/, since depth-first search
takes ‚.V C E/ time and it takes O.1/ time to insert each of the jV j vertices onto
Usefulness:  Groups  of  people  or  populations  that  are  connected  amongst  
the front of the linked list.
themselves;;  brain  regions  that  are  strongly  connected;;  figuring  out  if  a  network  
We prove the correctness of this algorithm using the following key lemma char-
such  as  telephone  lines  are  strongly  connected  (you  can  dial  anyone  in  the  
acterizing directed acyclic graphs.
network);;  Web  connectivity.  
 We  will  use  Depth  First  Search  (DFS),  but  not  just  by  itself.  

616
Consider  panel  (a)  in  the  figure  below  (we  will  look  at  panels  b  and  c  afterwards):  
Chapter 22 Elementary Graph Algorithms

a b c d
13/14 11/16 1/10 8/9

(a)

12/15 3/4 2/7 5/6

e f g h

a b c d

(b)

e f g h

cd

(c) abe

fg h

Figure 22.9 (a) A directed graph G. Each shaded region is a strongly connected component of G.
Each vertex is labeled with its discovery and finishing times in a depth-first search, and tree edges
are shaded. (b) The graph G T , the transpose of G, with the depth-first forest computed in line 3
of S TRONGLY-C ONNECTED -C OMPONENTS shown and tree edges shaded. Each strongly connected
component corresponds to one depth-first tree. Vertices b, c, g, and h, which are heavily shaded, are
the roots of the depth-first trees produced by the depth-first search of G T . (c) The acyclic component
graph G SCC obtained by contracting all edges within each strongly connected component of G so
that only a single vertex remains in each component.
 
Our algorithm for finding strongly connected components of a graph G D
Each  cluster  of  strongly  connected  components  is  in  gray.  For  example,  there  is  
.V; E/ uses the transpose of G, which we defined in Exercise 22.1-3 to be the
a  path  from  node  c  to  d,  and  from  node  d  to  c.  There  is  also  a  path  from  any  
graph G T D .V; E T /, where E T D f.u; !/ W .!; u/ 2 Eg. That is, E T consists of
node  in  a,b,e  to  other  nodes  in  that  cluster.  If  we  started  DFS  from  node  c,  we  
the edges of G with their directions reversed. Given an adjacency-list representa-
could  discover  strongly  connected  component  cluster  c  and  d.  But  if  we  started  
tion of G, the time to create G T is O.V C E/. It is interesting to observe that G
the  DFS  search  from  node  a,  we  will  actually  discover  all  the  nodes  of  this  graph,  
and G T have exactly the same strongly connected components: u and ! are reach-
and  not  an  individual  strongly  connected  component  cluster  (because  b  has  an  
able from each other in G if and only if they are reachable from each other in G T .
arrow  out  of  the  cluster  going  to  c,  so  c  would  be  discovered  too,  although  it  is  
Figure 22.9(b) shows the transpose of the graph in Figure 22.9(a), with the strongly
not  strongly  connected  with  a,b,e).  
connected components shaded.
Solution:  We  need  to  call  BFS  from  the  right  places  to  find  the  strongly  connected  
components!  
Algorithm:  
1.  Call  DFS  on  graph  G  to  compute  finish  times  u.f  for  each  vertex  u  in  graph  G  
(this  is  step  a  in  the  figure  above)  

2.  Compute  the  transpose  of  graph  G:   G T  

 
This  is  the  graph  G,  but  with  all  edge  directions  reversed  (see  panel  b  in  figure  
above).  Note  that  G  and  its  transpose  have  the  same  strongly  connected  
component  clusters.  
3.  Call  DFS  on  the  transpose  graph,  but  in  the  DFS  loop  consider  vertices  in  
decreasing  order  of  finish  time  from  the  DFS  we  ran  on  G  (starting  from  highest  
finish  time).  This  process  is  also  shown  in  panel  b.  Since  we  are  considering  
vertices  in  decreasing  order  of  finish  times  from  the  DFS  in  panel  (a),  we  first  
consider  vertex  b  and  perform  DFS  in  panel  (b)  on  the  transposed  graph.  We  find  
strongly  connected  components  b,a,e  in  this  way.  From  this  cluster  of  3  vertices,  
there  are  no    are  no  more  vertices  going  out  in  the  transpose  graph.  So  as  in  
DFS,  we  choose  a  new  starting  node,  and    again  take  the  largest  remaining  finish  
time  from  panel  (a)  and  therefore  start  in  panel  (b)  from  node  c.  We  find  nodes  c  
and  d  in  this  way.  Again,  there  are  no  more  nodes  going  out  of  the  cluster.  We  
start  from  g,  and  find  strongly  connected  component  cluster  g,f.  Finally,  we  also  
start  from  vertex  h  and  find  only  the  vertex  itself  as  the  last  strongly  connected  
component  (with  itself).  
4.  Panel  c  shows  the  resulting  strongly  connected  components,  with  each  cluster  
written  out  as  its  own  vertex.  This  also  shows  the  idea  of  strongly  connected  
components.  Each  cluster  has  a  path  between  all  nodes  in  the  cluster,  but  
between  clusters  there  is  only  a  single  arrow  (otherwise  if  there  were  arrows  
going  both  ways,  both  clusters  would  form  a  larger  strongly  connected  
component).  You  could  also  see  that  the  strongly  connected  components  of  
graph  G  and  its  transpose  are  the  same  (just  the  arrow  between  strongly  
connected  components  is  reversed  in  the  transpose  graph).  
Why  it  works:  The  main  intuition  is  that  the  finish  times  in  the  original  DFS  on  
graph  G,  give  us  the  proper  starting  points  for  our  second  DFS,  such  that  we  
each  time  discover  a  strongly  connected  component  cluster  and  there  are  no  
arrows  to  other  undiscovered  clusters.  For  example,  the  first  cluster  b,a,e  in  
panel  (b),  had  no  out  going  arrows  from  these  vertices.  The  second  cluster  c,d  
only  had  out  going  arrows  to  the  cluster  that  was  already  discovered.  And  so  on.  
There  is  a  lemma  and  proof  in  the  book,  which  we  did  not  go  over.