Scribd
Upload
33 views2 pages

Further Mathematical Methods (Linear Algebra) 2002 Problem Sheet 10

1. The document provides a problem sheet with exercises related to calculating inverses of matrices and proving theorems about them. It includes problems finding right and left inverses, showing a matrix has a right inverse but no left inverse, calculating a strong generalized inverse, and proving properties of inverses. 2. The optional problems provide additional exercises on finding singular value decompositions and using them to calculate generalized inverses, properties of matrices where the product is zero, and characterizing least squares solutions. 3. The harder optional problems ask to prove equivalences related to invertibility and properties of generalized inverses.

Uploaded by

Gag Paf
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
33 views2 pages

Further Mathematical Methods (Linear Algebra) 2002 Problem Sheet 10

1. The document provides a problem sheet with exercises related to calculating inverses of matrices and proving theorems about them. It includes problems finding right and left inverses, showing a matrix has a right inverse but no left inverse, calculating a strong generalized inverse, and proving properties of inverses. 2. The optional problems provide additional exercises on finding singular value decompositions and using them to calculate generalized inverses, properties of matrices where the product is zero, and characterizing least squares solutions. 3. The harder optional problems ask to prove equivalences related to invertibility and properties of generalized inverses.

Uploaded by

Gag Paf
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Further Mathematical Methods (Linear Algebra) 2002

Problem Sheet 10
(To be discussed in week 1 [Lent term] classes. Please submit answers to the asterisked
questions only.)

In this sheet, we try our hand at calculating some of the new kinds of inverses that we have introduced
in the lectures. We also prove and verify some theorems about them.

1. * Find the equations that u, v, w, x, y and z must satisfy so that



u v
1 1 1 1 0
w x = ,
1 1 2 0 1
y z
and hence find all right inverses of the matrix

1 1 1
A= .
1 1 2

Hence, verify that the matrix equation Ax = b has a solution for every vector b = [b1 , b2 ]t R2 .
What does the existence of many right inverses tell us about the number of solutions to the matrix
equation Ax = b for a given b R2 ? Explain this relationship in terms of the null-space of A.

2. Find the equations that u, v, w, x, y and z must satisfy so that



1 1
u v w 1 0
1 1 = ,
x y z 0 1
1 2
and hence find all left inverses of the matrix

1 1
A = 1 1 .
1 2
Using this, find the solutions of the matrix equation Ax = b for an arbitrary vector b = [b1 , b2 , b3 ]t
R3 . Hence, by considering the Cartesian equation of the plane that contains all of the vectors in
the range of A, or otherwise, verify that this matrix equation has a unique solution when it has a
solution.

3. * Show that the matrix


1 1 1
,
1 1 1
has a right inverse, but no left inverse. Further, find one of these right inverses. [There is no need to
find the general form of a right inverse as in Question 1, just find one particular right inverse.]

4. * Calculate the strong generalised inverse of the matrix



1 4 5 3
2 3 5 1
A= 3
,
2 5 1
4 1 5 3
by finding an m k matrix B and a k n matrix C such that the ranks of B and C are both k.

5. * Show that: If A has a right inverse, then R = At (AAt )1 is the strong generalised inverse of A.
Further, show that x = Rb is the solution of Ax = b nearest to the origin. [Hint: Start by showing
that R is a WGI.]
Other Problems. (These are not compulsory, they are not to be handed in, and they will not be
covered in classes.)

Here are some other questions on generalised inverses which you might like to try.

6. Find the singular values decomposition of the matrix



0 1
A = 1 0 ,
1 1
and hence calculate its strong generalised inverse. Also, find the orthogonal projections of R3 onto
R(A) and R2 parallel to N (A).

7. Given that the singular values decomposition of an m n matrix A is


k p
X
A= i xi yi ,
i=1

prove that the matrix given by


Xk
1
yi xi ,
i=1
i
is the strong generalised inverse of A.

8. Suppose that the real matrices A and B are such that ABt B = 0. Prove that ABt = 0.

9. Let A be an m n matrix. Show that the general least squares solution of the matrix equation
Ax = b is given by
x = AG b + (I AG A)z,
where z is any vector in Rn .

Harder Problems. (These are not compulsory, they are not to be handed in, and they will not be
covered in classes.)

Here are some results from the lectures on generalised inverses that you might like to try proving.

10. Prove that the following statements about an m n matrix A are equivalent:
1. A has a left inverse, i.e. there is a matrix L such that LA = I. (For example (At A)1 At .)
2. Ax = b has a unique solution when it has a solution.
3. A has rank n.

11. Prove that the following statements about an m n matrix A are equivalent:
1. A has a right inverse, i.e. there is a matrix R such that AR = I. (For example At (AAt )1 .)
2. Ax = b has a solution for every b.
3. A has rank m.

12. Prove that a matrix A has exactly one strong generalised inverse.

13. Consider an m n matrix A that has been decomposed into the product of an m k matrix B
and a k n matrix C such that the ranks of B and C are both k. Show that the matrix
AG = Ct (CCt )1 (Bt B)1 Bt ,
is a strong generalised inverse of A.

14. Suppose that Ax = b is an inconsistent set of equations: show that x = AG b is the least squares
solution that is closest to the origin. Is it necessary that this set of equations is inconsistent?

You might also like