pp. 247-248

  1. For each of the following matrices, determine a basis for each of the subspaces R(AT), N(A), R(A), and N(AT)

(a)

Change A to row-echelon form

Basis of R(AT) is the basis of row space of A which is .

N(A) is the null space of A which is the solution of

.

Hence, the basis of N(A) is

To find basis of R(A), and N(AT), reduce AT to reduce row-echelon form

Basis of R(A) is the basis of row space of AT which is .

N(AT) is the null space of A which is the solution of

.

Hence, the basis of N(AT) is .

(b)

To find the basis of R(AT) we have A to its reduced row-echelon form.

Bases of R(AT) is the basis of row space of A which are .and

N(A) is the null space of A which is the solution of

To find basis of R(A), and N(AT), reduce AT to reduce row-echelon form

Bases of R(A) is the basis of row space of AT which are .and

Since, the rank of A is 2 which is equal to number of column of 2, the dimension of N(AT) is equal to number of column minus number of rank = 2-2 =0. This means that N(AT) consists of only. Hence, N(AT) does not have basis.

(c)

To find the basis of R(AT) we have A to its reduced row-echelon form.

Bases of R(AT) is the basis of row space of A which are .and

Since, the rank of AT is 2 which is equal to number of column of 2, the dimension of N(A) is equal to number of column minus number of rank = 2-2 =0. This means that N(A) consists of only. Hence, N(A) does not have basis.

To find basis of R(A), and N(AT), reduce AT to reduce row-echelon form

Bases of R(A) is the basis of row space of AT which are .and

N(AT) is the null space of AT which is the solution of

.

Here, we have

,

which lead to

.

Hence,

(d)

To find the basis of R(AT) we have A to its reduced row-echelon form.

Bases of R(AT) is the basis of row space of A which are , .and

N(A) is the null space of A which is the solution of

which can be written in the form of augmented matrix as

Here, we have

,

and

which lead to

Hence,

To find basis of R(A), and N(AT), reduce AT to reduce row-echelon form

Bases of R(A) is the basis of row space of AT which are , .and

N(AT) is the null space of AT which is the solution of

.

Here, we have

,

and

which lead to

which lead to

Hence,

2. Let S be the subspace spanned by

Let A= be a matrix of size 1 by 3. We observe that R(AT), row space of A, is spanned by x. By Theorem 5.2.1, the .

Null space of A is determined by solving

From the above equation, we have

or

Hence, the basis of is and .

4. Let S be the subspace in spanned by , . Find a basis of

Let A= be a matrix of size 2 by 4. We observe that R(AT), row space of A, is spanned by x. By Theorem 5.2.1, the .

Null space of A is determined by solving

,

which can be written in the augmented form as

From the above equation we can conclude that

and

This leads to

Hence, the basis of is and .

pp. 258-259

1. Find the least square solution to each of the following systems.

(a)

and

We have that

and

.

The least square solution is determined by solving

,

which can be written in the augmented matrix form as

.

Hence, the solution is and

(b)

and

We have that

and

.

The least square solution is determined by solving

,

which can be written in the augmented matrix form as

.

Hence, the solution is and

(c)

We have that

and

.

The least square solution is determined by solving

,

which can be written in the augmented matrix form as

.

Hence, the solution is , and

2.

For (1.a), we have

and

.

Check that by determining. If then .

For (1.b), we have

and

.

Check that by determining . If then

For (1.c), we have

and

.

Check that by determining . If then

5.

(a) Find the best least square fit by a linear function to the data

Liner function implies that where a and b are to be determined by the data. Hence, we write

,

or

.

Obviously, the above equation is inconsistent. We use the least square approach to solve the problem.


and

The least square solution is determined by solving

,

which can be written in the augmented matrix form as

Hence, the solution is and

(b)

7. Given a collection of points (x1,y1)…(xn,yn) and Let

and

and let be the linear function that gives the best least square fit to the points.

We can write in the form of equations as

or in the form of matrix as

Obviously, the above equation is inconsistent. We use the least square approach to solve the problem.

and

Since we have

The least square solution is determined by solving

,

which can be written in the augmented matrix form as

Hence, the solution is and .

pp. 268

7.

(a)

(b)

(c)

8.

(a)

Hence,

(b) p is the projection of 1 onto x

Hence, we have

(c)

So, we have

.

9

Hence, are orthogonal

Since are orthogonal, the Pythagorean law holds. As a result, we have

Hence distance between two vectors is .

pp. 286-289

2

(a)

, ,

(b)

./

Hence, we have

and

9.

(a)

(b.i)

(b.ii)

(b.iii)

(b.iv)

26.

(a)

Hence, 1 and 2x-1 are orthogonal .

(b)

Hence , we have

Hence , we have .

(c)

We can approximate as