TEL 341-NETWORK ANALYSIS & COMPUTER AIDED DESIGN

Engr. T. J. Shima
MODULE 2: MATRIX REPRESENTATIONS OF A GRAPH
2.1 Introduction
Although a pictorial representation of a graph is very convenient for visual study, a matrix is a
more convenient and useful way of representing a graph to a computer. Therefore, in many
applications of graph theory, such as in electrical network analysis and operations research,
matrices turn out to be the natural way of expression the problem.
2.2 Incidence (or Node) Matrix of a Digraph
The incidence matrix is one of the matrices which represents a graph uniquely. It facilitates
the testing and identification of the independent variables.
For a given graph with 𝑛 nodes and 𝑏 branches, the complete incidence matrix 𝑨𝑖 is a
rectangular matrix of order 𝑛 × 𝑏, whose elements are given by:

1, if jth branch is associated with ith node and oriented away from ith node.

aij = −1, if jth branch is associated with ith node and oriented towards ith node. (2.1)
0, if jth branch is not associated with ith node.

This matrix, therefore, tells us which branches are incident at which nodes and what are the
orientations relative to the nodes.

Example 2.1
Consider the network of Figure 2.1.

Figure 2.1: Network for Example 2.1

Draw the graph of this network and obtain the incidence matrix for the graph.

Solution
The graph of the network in Figure 2.1 is as shown in Figure 2.2.

Figure 2.2: Graph of the network in Figure 2.1

1
 TEL 341-NETWORK ANALYSIS & COMPUTER AIDED DESIGN
The incidence matrix for the graph in Figure 2.2 is as follows:

2.2.1 Properties of Incidence Matrix

1. The sum of the entries of any column is zero.
2. The determinant of the incidence matrix of a closed loop is zero.
3. The rank of incidence matrix of a connected graph is (𝑛 − 1), where 𝑛 is the number
of nodes of the graph.
2.2.2 Number of Possible Trees of a Graph
The incidence matrix can be used to determine the number of possible trees of a graph. The
number of possible trees of a graph = 𝑑𝑒𝑡(𝑨𝑟 × 𝑨𝑇𝑟 ), where 𝑨𝑟 is the reduced incidence matrix
obtained by eliminating any one row of the complete incidence matrix 𝑨𝑖 , and 𝑨𝑇𝑟 is the
transpose of the matrix 𝑨𝑟 .

Example 2.2
Consider the graph shown in Figure 2.3.

Figure 2.3: Graph for Example 2.2

Determine the number of possible trees of the graph.

Solution
The incidence matrix for the graph in Figure 2.3 is as follows:

1 1 0 0 0 1
−1 0 1 0 −1 0
𝑨𝑖 = [ ]
0 −1 −1 1 0 0
0 0 0 −1 1 −1

By eliminating row 4, the reduced incidence matrix is obtained as:

1 1 0 0 0 1
𝑨𝑟 = [−1 0 1 0 −1 0]
0 −1 −1 1 0 0

2
 TEL 341-NETWORK ANALYSIS & COMPUTER AIDED DESIGN
Thus, the number of possible trees of the graph of Figure 2.3 is:

1 −1 0
1 0 −1
1 1 0 0 0 1 3 −1 −1
0 1 −1
𝑑𝑒𝑡 [−1 0 1 0 −1 0] × = |−1 3 −1| = 16.
0 0 1
0 −1 −1 1 0 0 −1 −1 3
0 −1 0
{ [1 0 0]}

2.3 Circuit Matrix of a Digraph

The circuit matrix 𝑩𝑐 of a directed graph is with 𝑒 branches and 𝑙 loops is an 𝑙 × 𝑒 matrix
whose elements are given by 𝑏𝑖𝑗 , such that:

1, if ith loop includes jth branch, and the orientations of the branch and the loop coincide.

bij = −1, if ith loop includes jth branch, but the orientations of the two are opposite.
0, if ith loop does not include the jth branch.


For example, consider the digraph given in Figure 2.4.

Figure 2.4: A digraph

The circuit matrix of the digraph in Figure 2.4 is:

Note that the orientations assigned to each of the four loops is entirely arbitrary. The loop in
the first row is assigned clockwise orientation, in the second row counterclockwise, the third
counterclockwise, and the fourth clockwise. Changing the orientation of any loop will simply
change the sign of every nonzero entry in the corresponding row.

3
 TEL 341-NETWORK ANALYSIS & COMPUTER AIDED DESIGN
2.4 Tie-Set Matrix

The fundamental circuits (or tie-sets) each made by a link (with respect to some specified
spanning tree) define a tie-set matrix or fundamental circuit matrix 𝑩𝑓 for a digraph. The
orientation assigned to each of the fundamental circuit is chosen to coincide with that of the
link.

For example, we can obtain the fundamental circuit matrix for the graph in Figure 2.4 by
selecting the spanning tree shown in Figure 2.5.

Figure 2.5: A tree and its cotree for Figure 2.4

The fundamental circuit matrix for the digraph in Figure 2.5 is as follows:

It may be observed here that 𝑩𝑓 = (𝑰𝜇 |𝑩𝒕 ), where 𝑰 is the identity matrix, 𝜇 is the number of
links, and 𝑡 corresponds to the number of twigs of the spanning tree.

2.5 Cut-Set Matrix

The cut-set matrix Qa =  qij  has one row for each possible cut-set of the graph and one
column for each edge, and is defined by:

1, if branch j is in cut-set i and the orientations agree,


qij = −1, if branch j is in cut-set i and the orientations are opposite,
0, if branch j is not in cut-set i.


For example, for the graph of Figure 2.6, there are seven cut-sets (one of which, consisting of
branches 𝑎, 𝑏, 𝑒, and 𝑓, is not shown), and so the cut-set matrix appears as:

4
 TEL 341-NETWORK ANALYSIS & COMPUTER AIDED DESIGN

Figure 2.6: Example graph for oriented cut-sets.

2.6 Fundamental Cut-Set Matrix

For a given graph, a fundamental cut-set matrix Q f is defined as a rectangular matrix whose
rows correspond to the fundamental cut-sets (number of twigs) and whose columns
correspond to the branches of the graph. Its elements are defined by

1, if branch j is in cut-set i and the orientations agree,


qij = −1, if branch j is in cut-set i and the orientations are opposite,
0, if branch j is not in cut-set i.

For example, consider the graph shown in Figure 2.7.

Figure 2.7: Fundamental cut-sets

With the tree selected as shown, the fundamental cut-sets are:

5
 TEL 341-NETWORK ANALYSIS & COMPUTER AIDED DESIGN
𝐶1 = {1, 2, 6}, 𝐶2 = {2, 3, 5, 6}, and 𝐶3 = {4, 5, 6}. Therefore, the fundamental cut-set
matrix is obtained as follows:
1 2 3 4 5 6
1 1 0 0 0 1
𝑸𝑓 =
[ 0 1 1 0 1 1]
0 0 0 1 −1 −1
2.7 Adjacency Matrix
Let 𝐺 be a digraph with 𝑛 nodes, containing no parallel branches. Then the adjacency matrix
𝑿𝑎 of the diagraph is an 𝑛 by 𝑛 matrix whose elements are given by:

1, if there is a branch directed from ith node to jth node,

xij = 
0, otherwise.
A diagraph is shown in Figure 2.8, its adjacency matrix is as shown.

Figure 2.8: A digraph

The adjacency matrix occurs in many different disciplines, and therefore has different names.
In the theory of sequential machines, it is called transition matrix. In the calculus of relations,
it is called the relation matrix. In network flows, it is called the connection matrix. It is also
known as the precedence matrix or preference in some sociological applications. In scheduling
and critical-path analysis, the adjacency matrix is known as the predecessor matrix.