Scribd
Upload
11 views2 pages

NUS MA2001 Linear Algebra I Tutorial 5

MA2001 tutorial

Uploaded by

hnklinh3205
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
11 views2 pages

NUS MA2001 Linear Algebra I Tutorial 5

MA2001 tutorial

Uploaded by

hnklinh3205
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

National University of Singapore

Department of Mathematics

Semester 1, 2025/26 MA2001 Linear Algebra I Tutorial 5

Questions in this tutorial sheet will be discussed in the tutorial classes in Week 7 (29/9-
3/10).
You are advised to revise Sections 3.3-3.4 before attempting the questions.

Questions

1. Any subset of Rn that can be written as a linear span is a subspace. Conversely,


every subspace can be written as a linear span. Determine, with justification,
which of the following are subspaces of R4 .

(a) {(0, 0, 0, 0)}.


(b) {(a, a + 1, a + 2, a + 3) | a ∈ R}.
(c) {(0, 0, 0, 0)} ∪ {(a, a + 1, a + 2, a + 3) | a ∈ R}.
(d) {(x, 2x, y, 3y) | x, y ∈ R}.
(e) {(x, y, z, w) | x − y + 2z − w = 0 and 2x + 2y − z + 2w = 0}.
(f) {x | Ax = Bx} where A, B are square matrices of order 4. (Here we write
vectors x as column vectors.)

(Learning objectives: To understand what is a subspace and what properties


subspaces must have.)

2. (All vectors in this question are written as column vectors.) Let A be an n × n


matrix. Define WA to be the subset {u ∈ Rn | Au = 2u} of Rn .

(a) Show that WA is a subspace of Rn .


 
2 0 2
(b) Let A = 0 1 −1. Write down the subspace WA explicitly.
0 0 1

(Learning objectives: To be able to interpret a subset that is described implicitly


and recognise it as a subspace.)

3. Determine whether each of the following sets below are linearly independent.

(a) {(1, −1, 1), (−2, 2, −2)}.


(b) {(1, −1, 1), (2, 2, 2)}.
(c) {(1, 1, −1), (−2, 2, 1), (4, 0, 3)}.

1
(d) {(1, −2, 1), (1, 2, 5), (−1, 1, −2)}.

(Learning objectives: Applying the definition of linear independence.)

4. (Vectors in this question are written as column vectors.) Let A be a non zero 3 × 3
matrix such that A2 = 0.

(a) Show that there exists a non zero vector v ∈ R3 such that Av ̸= 0.
(b) For such a v found in part (a), show that {v, Av} is a linearly independent
set.

(Hints: For part (a), try using a proof by contradiction. For part (b), apply the
definition of linear independence again!)

5. (For discussion) Let B be a n × m matrix where the rows of B are n linearly


independent vectors in Rm .

(a) Explain why R, the reduced row-echelon form of B must have no zero rows.
(b) Hence, conclude that for all i = 1, 2, . . . , n, Bx = ei , where

e1 = (1, 0, . . . , 0), e2 = (0, 1, 0, . . . , 0), · · · , en = (0, . . . , 0, 1)

are vectors in Rn , is always consistent.

(Learning objectives: The result above shows that there exists a m × n matrix
X such that BX = In . We say that B has a right inverse.)

Common questions

Powered by AI

Since \( Ax = Bx \) is equivalent to \( (A-B)x = 0 \), this condition outlines the kernel (null space) of the matrix \( A-B \). The kernel of any matrix is a subspace because it includes the zero vector (since \( (A-B)0 = 0 \)), is closed under addition \( (A-B)(x+y) = (A-B)x + (A-B)y = 0 + 0 = 0 \), and scalar multiplication \( (A-B)(cx) = c((A-B)x) = c0 = 0 \). Therefore, \( \{ x \mid Ax = Bx \} \) is a subspace .

The existence of \( BX = I_n \) is inferred because the rows of \( B \), being linearly independent, allow \( B \) to operate as an injective map onto \( \mathbb{R}^n \). By aligning \( B \) with transformations that convert it into its reduced row echelon form, \( BX = I_n \) ensures that \( X \) reconstructs the identity matrix from \( n \) linearly independent transformations, verifying the surjectivity of this combination. Thus \( B \) has a right inverse \( X \) due to its full column rank .

For a matrix \( A \) where \( A^2 = 0 \), it must have zero eigenvalues because \( A^2v = \lambda^2 v = 0 \) for \( v \neq 0 \). Since \( \lambda^2 = 0 \) implies \( \lambda = 0 \), all eigenvalues of \( A \) must be zero. Additionally, \( A \) is not diagonalizable due to its nilpotent nature and requires generalized eigenvectors to form a basis, reflecting its effect on the geometric multiplicity .

If \( A^2 = 0 \), then \( A \) is a nilpotent matrix. To show there exists a non-zero vector \( v \) such that \( Av \neq 0 \), use proof by contradiction: assume all vectors \( v \) satisfy \( Av = 0 \). Then, \( (Av)A = 0 \) implies \( A^2v = 0 \). However, since \( A^2 = 0 \), \( A \) must not be the zero matrix, contradicting the assumption that \( Av = 0 \) for all non-zero \( v \). Hence, such a \( v \) must exist .

Assume \( \{ v, Av \} \) is linearly dependent; then there exist scalars \( a \) and \( b \), not both zero, such that \( av + bAv = 0 \). If \( b \neq 0 \), then \( Av = -\frac{a}{b}v \). Applying \( A \) again gives \( A^2v = -\frac{a}{b}Av = -\frac{a}{b}(-\frac{a}{b})v \), leading to \( A^2v = 0 \), which agrees with \( A^2 = 0 \). However, this requires \( Av = 0 \) contradicting \( Av \neq 0 \). Thus, \( v \) and \( Av \) must be linearly independent .

Since the rows of \( B \) are linearly independent, they span an \( n \)-dimensional space. When \( B \) is transformed to its reduced row echelon form \( R \), the pivot positions (leading 1s) must be present in each row to maintain linear independence and span the same space. Zero rows would indicate dependence among the original rows, which contradicts the given linear independence .

The consistency of \( Bx = e_i \) for each standard basis vector \( e_i \) indicates the existence of a solution \( x \) for each basis vector. This means there exists a matrix \( X \) such that \( BX = I_n \), where \( I_n \) is the identity matrix. A matrix \( X \) fulfilling \( BX = I_n \) is a right inverse of \( B \), confirming that \( B \) has full row rank (linearly independent rows) and allows \( X \) to satisfy this condition .

To determine if \( \{(1, -1, 1), (2, 2, 2)\} \) is linearly independent, assume a linear combination \( c_1(1, -1, 1) + c_2(2, 2, 2) = (0, 0, 0) \). This forms the system of equations \( c_1 + 2c_2 = 0 \), \( -c_1 + 2c_2 = 0 \), and \( c_1 + 2c_2 = 0 \). Solving this system shows that \( c_1 = 0 \) and \( c_2 = 0 \) are the only solutions, indicating the set is linearly dependent as the equations are inconsistent unless \( c_1 \) and \( c_2 \) are zero .

A subset \( S \) of \( \mathbb{R}^n \) is a subspace if it includes the zero vector, is closed under vector addition, and closed under scalar multiplication. The subset \( \{(a, a+1, a+2, a+3) \mid a \in \mathbb{R} \} \) cannot be a subspace because it does not include the zero vector (0,0,0,0). This subset's vectors are of the form \( (a, a+1, a+2, a+3) \), which would require \(a = 0\) to produce the zero vector, \((0,0,0,0) = (0, 1, 2, 3)\), which is not satisfied .

To show \( W_A \) is a subspace, verify three conditions: it contains the zero vector, is closed under addition, and is closed under scalar multiplication. Firstly, \( A0 = 2(0) = 0 \), so \( 0 \in W_A \). For closure under addition, if \( u, v \in W_A \), then \( Au = 2u \) and \( Av = 2v \) imply \( A(u+v) = Au + Av = 2u + 2v = 2(u+v) \). For closure under scalar multiplication, if \( u \in W_A \) and \( c \) is a scalar, \( A(cu) = cAu = c(2u) = 2(cu) \). Hence, \( W_A \) is a subspace .

You might also like