Sunteți pe pagina 1din 3

ECO 134: Applied Mathematics

Sec : 3, 6

[Course Instructor: Saima Khan(Sai)]


SPRING 2011

Matrix Algebra: Does a Unique Solution Exist?


he question Does a unique solution exist may be asked in different ways. Essentially, the concepts you
use for your answer is always the same however the way in which you phrase your answer will differ based
on what the question is asking for.

Example 1:
You are given the following set of linear equations. Does a unique solution exist?

So the coefficient matrix here is:

For a unique solution to exist, the following conditions must hold:

Necessary Condition:

Sufficient Condition:

the coefficient matrix, A, must be


square

the rows of the coefficient matrix must be


linearly independent.

Hence the necessary condition holds

Since |A|0, the rows are linearly


independent

This implies that the equations we


are dealing with are functionally
independent and consistent

Thus there is a possibility


that unique solution may
exists

Hence now we are


confirmed that a unique
solution does exist.

Page

This implies that there are 2 equations


and 2 unknown variables

Since, A is a 2x2 matrix --> it is a square


matrix.

Example 2:

ECO 134: Applied Mathematics


Sec : 3, 6

[Course Instructor: Saima Khan(Sai)]


SPRING 2011

The same question may be asked in a slightly different way:


Q: Based on the rank that you just found, explain whether a unique solution exists or not?
In this case, recall that we are still working with the same set of linear equations presented above. So how is
the answer different? Well the concept is still the same to get a unique solution the same necessary and
sufficient conditions must hold. However, the difference is only in the way you phrase your answer. Note
that the addition/change that has been made is shown in italics and underlined form.

For a unique solution to exist, the following conditions must hold:

Sufficient Condition:
Necessary Condition:
the coefficient matrix, A, must be square

Since, A is a 2x2 matrix --> it is a square


matrix.
Hence the necessary condition holds

the rows of the coefficient matrix must be


linearly independent, i.e., there must be full
rank

We find that rank of A=2, i.e, there are 2 lineraly


independent rows

This implies that there are 2 equations and


2 unknown variables

This implies that the equations we are


dealing with are functionally
independent and consistent

Thus there is a possibility


that unique solution may
exists

Hence now we are


confirmed that a unique
solution does exist.

Note that we used the term full rank in our answer what do we mean by full rank?
Well rank is the number of independent rows of columns that a matrix has. When we say that a matrix has
full rank, it means all the rows/columns are independent.

So what about a 2x3 matrix ?

Page

So, if we are dealing with a 2x2 matrix, if its rank is 2, then we say it is a full rank matrix. If we are dealing
with a 3x3 matrix, then if its rank is 3, we say it is a full rank matrix.

[Course Instructor: Saima Khan(Sai)]


SPRING 2011

ECO 134: Applied Mathematics


Sec : 3, 6
Q: Say the coefficient matrix is

where the rank is 2; in this case does a unique solution exist?

Ans:
In this case we can easily say that a unique solution does not exist because the necessary condition
does not hold, i.e., is a 2x3 matrix it is NOT SQUARE.
Here there are two equations but three unknown variables thus this system of linear equation
cannot be solved!

EXCEPTIONAL CASE: What then about a 3x2 matrix?


Q: Say the coefficient matrix is

where the rank is 2; in this case does a unique solution exist?

Ans:
This is an exceptional case. Even though the necessary condition does not hold, provided the rank is
2, we may get a solution. Note that here we have 3 equations but 2 unknown variables. If one of the
equations disappear ( because they are linearly dependent), then we are left with two linearly
independent rows.
This means we end up with two equations and 2 unknown variables there is a possibility of solving
this system of linear equation.
Say, it is the 3rd row the is redundant because it can be transformed into row 2. In that case we can
write a new matrix

if the

, then we can find a unique solution.

Nonsingularity:
Notice however, that for a 3x2 matrix, even though the unique solution may exist if the rank is 2, a
determinant will NEVER exist!
For a determinant to exist :
o The matrix MUST be square
o The rows must be linearly independent, i.e. ,
I see that a few of you often get confused between the terms singular and non-singular. Lets try to
simplify the concepts in playing a silly word game:
Matrix A is singular it means is single does not have a date it is all alone in the
world it does not have a pair,
, to spend its life with.
Matrix is non-singular it means is not single it has a date it has a partner, i.e. , an
to share its life with.
Properties of Determinant:

Page

Though we did not discuss the properties of a determinant, the above property is one property you
must keep in mind for your midterm.

S-ar putea să vă placă și