Anjan Gupta

LECTURE - 12
ANJAN GUPTA
Lemma 1. Let A be a row reduced echelon matrix of size n × n. If A is nonsingular, then A = In .

Proof. Since A is nonsingular, the last row of A must be a nonzero row vector (row vector with at
least one nonzero entry). By the property of row reduced echelon matrices, it follows that all rows
of A are nonzero. Let the leading 1 of the i-th row sits in the ki -th column for i = 1, 2, . . . , n. Then
we have
1 ≤ k1 < k2 < . . . < kn ≤ n.
This forces ki = i for i = 1, 2, . . . , n. Therefore, all diagonal entries are leading one’s of different
nonzero rows. This implies that all entries lying outside the diagonal are zero. Hence A = In .
Proposition 2. Let A be a square matrix of size n. Then A is invertible if and only A can be written
as a product of elementary matrices of type Ei (c) and Ei j (c), c being a scalar.
Proof. Elementary matrices are invertible. We also know that product of two invertible matrices is
invertible. Therefore, if A is a product of elementary matrices, then A is an invertible matrix.
Conversely, we assume that A is an invertible matrix. We apply a finite number of elementary
row operations of type Ri = R0i + cR j followed by a finite number of elementary row operation
of type R0i = cRi to obtain a row reduced echelon matrix R. Consequently, we obtain elementary
matrices E1 , E2 , . . . , Em of type Ei j (c) and elementary matrices E10 , . . . , En0 of type Ei (c) such that
En0 . . . E10 Em . . . E1 A = R.
Since A, Ei , E 0j are nonsingular matrices, the matrix R is also nonsingular. This implies that
R = In , so En0 . . . E10 Em . . . E1 A = In . It follows that A = E1−1 . . . Em−1 E10−1 . . . En0−1 . Now we know that
inverse of an elementary matrix is an elementary matrix of the same kind. Therefore A is a product
of elementary matrices.
Lemma 3. Let A be a square matrix. Then A = EDR where E is a product of elementary matrices
of type Ei j (c), D is an invertible diagonal matrix and R is a row reduced echelon matrix.
Proof. The converse part of the previous proposition shows that there are elementary matrices
E1 , E2 , . . . , Em of type Ei j (c) and elementary matrices E10 , . . . , En0 of type Ei (c) such that
En0 . . . E10 Em . . . E1 A = R
is a row reduced echelon matrix. It follows that A = E1−1 . . . Em−1 E10−1 . . . En0−1 R. Now each of E 0−1
j is
an elementary matrix of type Ei (c) which is an invertible diagonal matrix. Therefore their product
E10−1 . . . En0−1 is equal to an invertible diagonal matrix D.
Each Ei−1 is an elementary matrix of type Ei j (c). Let E = E1−1 . . . Em−1 , then E is a product of
elementary matrices of type Ei j (c). Now we observe that A = EDR. Hence the result follows.
Date: 28/01/2020.
1
LECTURE - 12 ANJAN GUPTA
Theorem 4. Let A, B be matrices of the same size. Then det(AB) = det A det B.
Proof. First we assume that A is a diagonal matrix of size n. If diagonal entries of A are d1 , . . . , dn ,
then AB is obtained from B by multiplying rows 1, 2, . . . , n by scalars d1 , d2 , . . . , dn respectively.
This implies det(AB) = d1 . . . dn det B. But det A = d1 . . . dn . Therefore det(AB) = det A det B.
If A is a product of elementary matrices of type Ei j (c), then det A = 1. The product AB can be
obtained from B by applying a finite sequence of elementary operations of type R0i = Ri + cR j on B,
so det(AB) = det B. Therefore det(AB) = det A det B.
Now we prove the result in the general case. By Lemma 3, we obtain A = EDR where E is a
product of elementary matrices of type Ei j (c), D is an invertible diagonal matrix and R is a row
reduced echelon matrix. We consider two cases
Case - 1 : A is a singular matrix.
Since A = EDR and E, D are nonsingular matrices, R is a singular row reduced echelon matrix.
This implies that the last row of R is a zero row. Consequently, the last row of RB is also zero. It
follows that det(RB) = 0. Now
det(AB) = det(EDRB) = det(DRB) = det D det(RB) = 0.
In this case det A = 0. Therefore, det(AB) = det A det B.
Case - 2 : A is a nonsingular matrix. In this case R is a nonsingular row reduced echelon matrix.
By Lemma 1, R = In . Consequently A = ED. Therefore det(A) = det(D). Now
det(AB) = det(EDB) = det(DB) = det D det B = det A det B.
Hence the proof is complete.
Department of Mathematics, Indian Institute of Science Education and Research Bhopal, Bhopal Bypass Road,
Bhopal, Madhya Pradesh, India. Pin - 462066.
Email address: [email protected]

Anjan Gupta

Uploaded by

Copyright:

Available Formats

Anjan Gupta

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Anjan Gupta

Uploaded by

Copyright:

Available Formats

LECTURE - 12

Lemma 1. Let A be a row reduced echelon matrix of size n × n. If A is nonsingular, then A = In .

You might also like