Lecture 12: Graphs
Graph Terminology
2
1
3
vertex, node, point
edge, line, arc
G = (V, E)
V is set of vertices
E is set of edges
Each edge joins two different vertices
1
2
Undirected Graph
edges do not have a direction
The edge from 1 to 2 is also an edge
from 2 to 1
Edge (1, 2) implies that there is also
an edge (2, 1) [ The same edge ]
3
1
2
Directed Graph
edges have a direction
Edge (2, 1) means only that there is an
edge from 2 to 1
In this example,there is no edge from 1
to 2
4
12
2
22
15
1
3
13
32
16
Weighted Graph
weights (values) are associated with the
edges in the graph
may be directed for undirected
Weighted graphs are also referred to as
networks
5
1
2
Complete Graph
For each pair of vertices, there is one
edge
If G = (V, E) is a complete graph, and
|V| = n, then can you calculate |E|?
1
2
Subgraph
A subgraph G of graph G = (V, E) is a
graph (V, E) that V V and E E.
Path
the sequence of edges (i1, i2), (i2,
i3), ,(ik-1, ik).
Denoted as path i1, i2, ..., ik
Simple path all vertices (expect
possibly first and last) are different
Length of path is sum of the lengths of
the edges
Examples of graph
Representation of Graphs
Adjacency matrix
Incidence matrix
Adjacency lists: Table, Linked List
Space/time trade-offs depending on
which operation are most frequent as
well as properties of the graph
10
11
Can we use tree traversal
algorithms to traverse graphs?
Why?
Loops in graphs, parent/child
relation in trees
12
Depth first search
DFS(v)
num(v)=i++;
for all vertices u adjacent to v
if num(u) is 0
attach edge(uv) to edges;
DFS(u);
depthFirstSearch()
for all vertices v
num(v)=0;
edges=null; i=1;
while there is a vertex v such that num(v) is 0
DFS(v);
output edges
13
14
15
Breadth first search
breadthFirstSearch()
for all vertices u
num(u)=0;
Edges=null; i=1;
While there is a vertex v such that num(v)==0
num(V)=i++;
enqueue(v);
while queue is not empty
v=dequeue();
for all vertices u adjacent to v
if num(u) is 0
num(u)=i++;
enqueue(u);
attach edge(uv) to edges;
Output edges;
16
17
18
An Example
1
What would the visit orders for
DFS(1), DFS(5), BFS(1), BFS(5)
look like?
19
Unweighted Shortest Path
Find the shortest path (measured by
number of edges) from a designated
vertex S to every vertex
Simplified case of weighted shortest
path
20
Algorithm
Starting from node S
Distance from S to S is 0, so label as 0
Find all nodes which are distance 1 from S
Label as distance 1
Find all nodes which are distance 2 from S
These are 1 step from those labeled 1
This is precisely a breadth first search
21
An Example
3
1
2
4
5
7
What are the values of distance[2], distance[3], ,
distance[7] after FindShortestPath(G, 1) is executed?
22
Positive Weighted Shortest Path
Length is sum of the edges costs on the
path
All edges have nonnegative cost
Find shortest paths from some start
vertex to all vertices
similar process to unweighted case
Dijkstra's Algorithm(Single source
shortest path)
23
Distance at each node v is shortest path
distance from s to v using only known
vertices as intermediates
An example of a Greedy Algorithm
Solve problem in stages by doing what
appears to be the best thing at each
stage
Decision in one stage is not changed
later
A simple greedy example is counting
change (quarters, then dimes,)
24
Greedy algorithms do not always work
Change breaks down with addition
of .12 coin ( try .15 value )
They do work for certain problems and
are fairly simple
25
DijkstraAlgorithm(weighted Simple digraph, vertex
first)
for all vertices v
currDist(v)=infinity;
currDist(first)=0;
toBeChecked=all vertices;
While toBeChecked is not empty
v = a vertex in toBeChecked with minimal
currDist(v);
remove v from toBeChecked;
for all vertices u adjacent to v and in
toBeChecked
if currDist(u)>currDist(v)+weight(edge(uv))
currDist(u)=currDist(v)+weight(edge(uv));
predecessor(u)=v;
26
start at v1, all distances are infinity
8
5
2
6
mark v1 as known, with distance 0
0
2
1
2
4
3
1
2
2
3
4
8
5
4
1
10
10
5
4
1
6
27
0
1
2
3
2
4
8
5
10
4
1
Now adjust distances for v2 and v4
0
1
2
2
2
1
4
8
5
10
4
1
6
28
Select v4 as known, since it has the lowest cost among unknown
0
1
2
2
2
1
4
8
5
10
4
1
Examine adjacent, adjust v3, v5, v6 and v7
0
2
2
1
2
10
4
3
1 1
2
2
4
5 3
3 3
8
4
5
6
1
6
7
9
5
29
Select v2 as known
0
1
3 3
2
2
1
4
10
5 3
6
1
6
7
9
5
Check v4, v5. v4 is final, v5 has lower cost of 3 vs. 12
0
2
2
1
2
10
4
3
1 1
2
2
4
5 3
3 3
8
4
5
6
1
6
7
9
5
30
Select v5 as known
0
1
3 3
2
2
2
1
4
1
6
9
Check v7, but not adjusted
0
1
2
2
1
4
8
5
6
9
5 3
6
7
5
3 3
10
10
5 3
4
1
7
5
6
31
v3 is selected as known
0
1
3 3
2
2
1
4
10
5 3
6
1
6
7
9
5
Check v1 and v6. v1 known, v6 adjusted from 9 to 8
0
2
2
1
2
10
4
1
3
1
2
2
4
5 3
3 3
8
4
5
6
1
6
7
8
5
32
Select v7 as known
0
1
4
1
2
3 3
2
2
2
1
4
8
5
6
8
10
5 3
4
1
Check v6, change from 8 to 6
0
2
1
4
1
3
1
2
4
3 3
8
4
5
1
6
6
7
5
2
2
10
2
7
5
5 3
6
33
Select v6 as known
0
1
4
1
2
3 3
2
2
2
1
4
8
1
Done
6
6
5 3
2
2
2
1
7
5
1
4
8
5
6
6
Nothing to check
0
1
4
2
3 3
10
10
5 3
4
1
7
5
6
34
Spanning tree
subgraph of G
contains all vertices of G
connected graph with no cycles
35
Examples of spanning trees
36
Minimum spanning tree
spanning tree with minimum cost
only exists if G is connected
number of edges is |V|-1
three greedy methods
Kruskal's algorithm
Dijkstras method
Prim's algorithm
differ in how next edge is selected
37
Kruskal's algorithm
KruskalAlgorithm( weighted connected
undirected graph)
tree = null;
edges = sequence of all edges of graph
sorted by weight;
for(i=1;i<= |E| and |tree| < |V|-1; i++)
if e i from edges does not form a cycle with
edges in tree
add e i to tree;
38
39
Construct MST for this graph using Kruskal's algorithm starting
v1
2
1
2
10
4
3
1
2
7
3
4
5
8
4
5
6
1
6
7
8
5
2
3
10
4
1
6
40
Construct MST for this graph using Kruskal's algorithm starting
v1
2
1
2
10
4
3
1
2
7
3
4
5
8
4
5
6
1
6
7
8
5
2
3
10
4
1
6
41
Construct MST for this graph using Kruskal's algorithm starting
v1
2
1
2
10
4
3
1
2
7
3
4
5
8
4
5
6
1
6
7
8
5
2
3
10
4
1
6
42
Construct MST for this graph using Kruskal's algorithm starting
v1
8
5
2
3
10
4
1
6
43
Dijkstras method
DijkstraMethod( weighted connected undirected
graph)
tree = null;
edges = an unsorted sequence of all edges of
graph;
for j=1 to |E|
add e i to tree;
if there is a cycle in tree
remove an edge with maximum
weight from this only cycle;
44
45
Prim's algorithm
Similar to the Dijkstras algorithm in
shortest paths
grow the tree in successive stages
in each stage, one node is picked as the
root, we add an edge, and thus a vertex
is added to the tree
have a set on vertices in the tree and a
set that is not in the tree
46
Prim's algorithm (cont.)
at each stage, a new vertex to add to
the tree is selected by choosing edge (u,
v) such that the cost of (u,v) is the
smallest among all edges where u is in
the tree and v is not
Build spanning tree starting from v1
Result in the same spanning tree as
that given by the Kruskal algorithm
47
Construct MST from v1 for this graph using Prim's algorithm
2
8
5
4
8
4
1
10
10
4
1
6
48
Construct MST from v1 for this graph using Prim's algorithm
2
8
5
4
8
4
1
10
10
4
1
6
49
Construct MST from v1 for this graph using Prim's algorithm
2
8
5
4
8
4
1
10
10
4
1
6
50
Construct MST from v1 for this graph using Prim's algorithm
Same MST as Kruskal
4
8
10
4
1
6
51
