APPLICATION OF LINEAR

ALGEBRA IN OTHER DISCIPLINES:

RISE OF INCIDENCE MATRIX.

by Himanshu Hani[MS23137]

November,2023
ABSTRACT

In this assignment,I have tried to demonstrate the application part of linear

algebra spread all over other disciplines.Actually one can wonder from where a
matrix came into picture.Actually there is a definite structure and anyone who
work with it,uses that structure.In this paper, i will try to demonstrate the appli-
cation of linear algebra in other field like in kirchoff’s law(electrodynamics),etc.

1
INTRODUCTION

LET’S DEFINE SOME USUAL NOTATION FOR THE READERS.:

(i).GRAPH: the path which consists of nodes and edges.
(ii).INCIDENCE MATRIX: it is a special matrix which helps to present the
circuit(a type of graph) in some definite structure which is known as matrix
nowadays.
(iii).Nodes: call it as a point of intersection of two path in a circuit or a
graph
This leads me to make your attention towards this incidence matrix. Actu-
ally it is just like adjacency matrix and both have derived from graph theory
itself.But in this text, i will try to stick on only incidence matrix and will discuss
more of its properties and its relation with other disciplines.

2
METHODOLOGY

Let’s consider a graph(basically call it an electric circuit) [see(fig.1)]. In this

circuit, we can observe there is 4 nodes( intersection point of two different
wires) and there is 5 edges. Now let’s play with this graph mathematically and
derived a language which is fundamental thing required to understand other
disciplines(like physics here) very quickly.
Now,I will try to orgnise all the information which i am observing from this
figure.But question is arising HOW WE WILL DO THAT?. So,basically here
linear algebra play it’s role.We will use some structured form for it ( later which
will become something, which is known as matrix).
First of all, let’s again look at the picture and try to understand it. we can
observe that in the whole circuit some amount of current is flowing from one
node to another. Like as e.g. there is a flow of current from node1 to node2
along with edge1 .
So i am defining some rules for the game:
(i). When current(shown its direction by arrow) flows from one node, we will
assign it a value of -1 and where its current converge ( along the same wire(or
same edge)) we will assign it a value of +1.
(ii). Apart from these two nodes all nodes which are not accountable for
these things got assigned value of zero.
(iii).If three rows becomes linearly dependent, we will call it a loop.
Now we are all set to proceed.
We will arrange the data in such a way so that its nodes makes columns
and itsedges makes rows of that tabular form ( which is known as ”incidence
matrix” nowadays but currently i don’t know what a matrix is).
So after the arrangement of assigned values according to the rule,we get
some unique structured format:
As we can determine it has some tabular structure which consist of rows
and columns.
Now observe carefully you will find that we have 4 columns in this matrix
which is reflecting the information of nodes and its 5 rows reflecting the path
along its five edges.
The image in next page reflects the format of matrix: (see fig.2)
Now,let’s take x1,x2,x3,x4 be the potential difference of respective nodes.
Now we can verify the potential difference between any two nodes simply by
solving system of homogeneous linear equation i.e. AX = 0

3
 Figure 1: a type of circuit(graph)

Figure 2: structured form of assign values(incidence matrix)

4
  
−1 1 0 0

 0 −1 1 0 

A=
 −1 0 1 0 

 −1 0 0 1 
0 0 −1 1
. This is our required incidence matrix.
Now let’s perform our calculation upon this. We want to solve:
AX = 0

 
−1 1 0 0  
 0 −1 1 0  x1
  x2 
AX =   −1 0 1 0   x3  = 0
 
 −1 0 0 1 
x4
0 0 −1 1
implies that:
x2 - x1 = 0
x3 - x2 = 0
x3 - x1 = 0
x4 - x1 = 0
x4 - x3 = 0
which simply denoting that between two nodes there is no potential difference
and we can also see that in our circuit there is notelectrical
 components and
1
1
thus there is no dissipation of energy. Thus, X =  1.C; where C belongs to


1
some Real number.
This X is in nullspace of the above system of linear equation.
JUSTIFICATION FOR LOOP IN THE CIRCUIT :
We can observe from fig.1 that for rows 1,2 and 3, their elements are linearly
dependent. In other words, if you add row1 and row2, you will simply obtain
row3. This shows that linear relation of these rows leads to non trivial solution.
So, we can conclude that.

LINEAR DEPENDENCY OF THREE ROWS OF INCIDENCE MATRIX LEADS TO ” LOOP

(1)
P.
Applying the same on our graph we can find that, Yes! there is a loop made
with nodes 1,2 and 3. 
1
1
Particularly, X = 
1 is a nullspace of A.


5
 Figure 3: when node 4 got earthed.

 
1
1
Then its basis will be : 1.


1
So, as we can notice that number of vector in basis is one. Therefore, di-
mension of this nullspace is one.
LET’S TAKE ANOTHER CASE :
Ground the node 4 of the circuit,that means x4 = 0.(see fig,3)
Then our
 system of equation
 got reduced to:
−1 1 0  
 0 −1 1  x1
  x2 
AX =   −1 0 1   x3  (1) = 0
 
 −1 0 0 
x4
0 0 −1
actually, the 4th column get vanish due to the grounding of node 4.
Now, the rank of this matrix will be = 3 Thus r = 3

6
PROOF OF KIRCHOFF’S
JUNCTION RULE USING
INCIDENCE MATRIX

Let’s recall kircoff’s junction rule for a while, it states that with respect to a
specific junction, net incoming and outgoing current is equal to zero.
I will try to prove it using, above learned concepts about incidence matrix.
previously,we are dealing with incidence matrix A,this time we will take
transpose of matrix A so that matrix order  will
 becomes 4x5 instead of 5x4.
  y1
−1 0 −1 −1 0 y2 
T
 1 −1 0 0 0   
y3  = 0
A X =  
0 1 1 0 −1   y4 

0 0 0 1 1
y5
implies that: −y1 - y3 - y4 = 0 y1 - y2 = 0 y2 + y3 - y5 = 0 y4 + y5 = 0
Now if you are wondering why there is only 4 equation instead of 5 as total
variable is 5. so, i want to make recall that there is 4x5 matrix that means it
has 5 variables but only 4 solution, that means one variable is free variable.
Thus the above linear equatios is proving kirchoff’s law as mentioned.
From (figure.4),we can get the feel of kirchoff rule, that for node1, currents
y1,y3 and y4 are leaving that’s why according to the convention of this rule
leaving currents have a negative sign, so that algebraic sum of all currents at
junction is 0.
That’s why we can see here, y1 - y3 - y4 = 0 [ for node 1]

7
Figure 4: for proving junction rule

8
CONCLUSION

So, in this assignment, i have tried my best to say whatever i can construct so
solidify the application part of linear algebra. Not only in physics, it plays a
vital role in chemistry,computer science also.
We can also say that linear algebra provide a medium with the help of that
programmers can program various method for circuit analyzer in electronics
field.
Readers are requested to try this program once and verify by yourself, what
differences you have observed.
// Java program to convert adjacency matrix // to incidence matrix
import java.io.*;
class GFG public int[][] adjacencyMatToIncidenceMat(int[][] adj) int ver-
tices = adj.length, edges = 0;
// count number of edges in the graph for (int i = 0; i ¡ adj.length; i++)
for (int j = i + 1; j ¡ adj[i].length; j++) if (adj[i][j] ¿ 0) edges++;
// construct incidence matrix int[][] incidenceMat = new int[adj.length][edges];
for (int i = 0; i ¡ adj.length; i++)
for (int j = i + 1; j ¡ adj[i].length; j++) int edgeNumber = adj[i][j];
if (edgeNumber ¿ 0) incidenceMat[i][edgeNumber - 1] = 1; incidenceMat[j][edgeNumber
- 1] = 1;
return incidenceMat;
// driver code public static void main(String[] args) GFG gfg = new GFG();
int[][] adj = 0, 1, 0, 4 , 1, 0, 2, 0 , 0, 2, 0, 3 , 4, 0, 3, 0 , ; int[][] incidence =
gfg.adjacencyMatToIncidenceMat(adj);
for (int[] row : incidence) for (int val : row) System.out.print(val); Sys-
tem.out.println();
THANK YOU