Applied Linear Algebra: Problem set-2

Instructor: Dwaipayan Mukherjee∗

Definitions: (1) For a non-empty set S and a field F, we define F[S; F] as F[S; F] := {f : S → F}, which is a
vector space over F.
(2) The vector space of all polynomials of degree less than or equal to n, with coefficients in the field F, is denoted
by F[x]n .

1. Consider the subspace of all quadratic polynomials over R, given by R[x]. Is the set G := {x2 + 1, x2 − 1, 2x +
3, x + 3, 5} a generating set for R[x]? Is G linearly independent? Justify your answers.
2. If a vector space V over Z2 contains linearly independent vectors x, y, z, then are x + y, y + z, z + x also linearly
independent? Justify your answer mathematically.
3. If a vector space V over a subfield of C contains linearly independent vectors x, y, z, then show that x + y, y +
z, z + x are also linearly independent.
4. Considering C, the set of complex numbers, to be a vector space over R, show that ‘1 and x are linearly
independent’ ⇐⇒ ‘x is not a real number’.
5. Are the vectors v1 = [1 1 2 4]T , v2 = [2 − 1 − 5 2]T , v3 = [1 − 1 − 4 3]T , v4 = [2 1 1 6]T linearly independent
in R4 ? Find a basis for the subspace ⟨{v1 , v2 , v3 , v4 }⟩.
6. Find a basis for symmetric 3 × 3 matrices, having real entries, as a vector space over R.
7. Consider the vector space F2×2 (F is any field). Obtain a basis for this vector space, say B = {A1 , A2 , A3 , A4 },
such that A2i = Ai , i = 1, 2, 3, 4. Establish that your answer is indeed a valid basis.
8. Suppose {v1 , v2 , v3 } is a basis for the vector space V over the the field F. Is the set {v1 + v2 + v3 , v2 + v3 , v3 }
also a basis for V? Justify your answer mathematically.
9. Prove that if {v1 , v2 , . . . , vn−1 , vn } is a basis for the finite dimensional vector space V, then so is the set
{v1 − v2 , v2 − v3 , . . . , vn−1 − vn , vn }.
10. Delineate a constructive method for obtaining the basis of W1 ∩ W2 when the bases for the finite dimensional
vector spaces W1 and W2 are given.
2×2
11. In the vector
 space F over F, consider the following subspaces: W1 ⊆ F2×2 such that it contains
 matrices
 of
α −α 2×2 p q
the form with α, β, γ ∈ F and W2 ⊆ F such that it contains matrices of the form with
β γ −p r
p, q, r ∈ F. What are the dimensions of W1 , W2 , W1 ∩ W2 , and W1 + W2 ?
12. For A ∈ Fm×n , B ∈ Fn×p , and C = AB, prove that rank(C) ≤ min(rank(A), rank(B)).
13. For the vectors v1 = [1 1 2 4]T , v2 = [2 −1 −5 2]T , v3 = [1 −1 −4 0]T , consider the subspaces W1 := span(v1 , v2 )
and W2 := span(v2 , v3 ).
(a) Obtain a basis for W1 ∩ W2 .
(b) Compute dim(W1 + W2 ).
14. Consider W := {p ∈ F[x]3 : p(x) = a0 + a1 x + a2 x2 + (a0 + a1 − a2 )x3 , a0 , a1 , a2 ∈ F}. Show that W is a
subspace of F[x]3 . Obtain a basis for W and hence evaluate dim(W). Suppose p1 (x) = 1 + x + x2 + x3 and
p2 (x) = 1 + x2 . Show that S = {p1 , p2 } is a linearly independent set in W. Extend S to a basis for W.
15. For any finite dimensional vector space, V, show that there exist subspaces W1 and W2 of V such that V =
W1 ⊕ W2 .
16. For A, B ∈ Rn×n , we have rank(A + B) ≤ rank(A)+rank(B). Prove or disprove the assertion.
space, V, suppose there exist subspaces Wi , i ∈ {1, 2, . . . , m} such that V =
17. For a finite dimensional vectorP
m
⊕m
i=1 W i . Show that dim(V) = i=1 dim(Wi ).

18. Show that ⟨{1, cos x, sin x}⟩ and ⟨{1, eix , e−ix }⟩ over C are the same vector space, say V. Also show that
{1, cos x, sin x} and {1, eix , e−ix } are both bases for V. Obtain the basis transformation matrix between them.
∗ Asst. Professor, Electrical Engineering, Office: EE 214D, e-mail: dm@ee.iitb.ac.in.

1
 19. Provide an example of each of the following: (a) an injective map from R2 to R2 , and (b) a surjective map from
R2 to R2 . Can you cook up similar examples from R2 to R3 ? Is it possible to do so from R3 to R2 ? Justify if
these are not possible, or provide examples if they are possible.
20. Suppose ϕ : V → W is a linear map between two finite dimensional vector spaces. Prove that “ϕ is injective” =⇒
dim(V) ≤ dim(W) and “ϕ is surjective” =⇒ dim(V) ≥ dim(W).
21. Consider a 3 × 3 matrix of real entries, say A ∈ R3×3 which is used to define a linear transformation ϕ :
R3×3 → R3×3 such that ϕ(X) := AX. Describe clearly, with suitable mathematical justifications, how you will
determine the rank of ϕ.

22. For a linear map φ : R2 → R2 that sends the vector (x1 , x2 ) to (−x2 , x1 ), show that φ2 = −identity. Obtain
[φ]B , where B = {(1, 2), (−1, 1)}. Further, show that  λ ∈ R, (φ − λ · identity) is invertible. Can you
 for any
′ 2 0 1
obtain a unique basis, B for R , such that [φ]B′ = ? Justify with proper calculations.
−1 0

23. Suppose T : R3 → R2 be a linear map such that T (p1 , p2 , p3 ) = (p1 − p2 , 5p3 − 2p2 ). If B and B ′ are the
standard bases in R3 and R2 , respectively, obtain the matrix representation of T relative to B and B ′ . Suppose
instead someone chooses B = {(−1, 0, 1), (1, −1, 1), (1, 0, 0)} and B ′ = {(1, 1), (0, 1)}, how would (s)he rewrite
the same matrix representation?
24. Suppose V is the vector space of all polynomials having real coefficients with degree less than or equal to
n < ∞. Let B = {f0 , f1 , . . . , fn } and B ′ = {g0 , g1 , . . . , gn } be two ordered bases for V such that fi = xi and
gi = (x + α)i for some scalar α. Represent the differentiation operator, D : V → V in both these ordered bases,
using basis transformation. What do you observe?
25. Consider the same vector space as in Qn. 24 with the ordered basis B, and n = 3. Consider the linear operator
d
D̂ : V → V given by D̂(f (x)) = 3f (x) + (3 − x) dx (f (x)). Obtain bases for Ker(D̂) and Im(D̂). Suppose
g(x) = 7 + 8x. What are all possible solutions of D̂(f (x)) = g(x)?

26. For vector spaces U, V, and W, of which U, and V are finite dimensional, consider two linear transformations
τ : U → V and ψ : V → W. Suppose the composition of the two transformations is given by ψ ◦ τ . Establish
that dim(Ker(ψ ◦ τ )) ≤ dim(Ker(τ )) + dim(Ker(ψ)). When does equality hold? What can you say about the
inequality when τ is surjective?

27. (Cryptography: Hill cipher ) A common way of encoding messages is to associate distinct positive integers with
each letter of the alphabet. For instance, a very basic encoding would be to choose A → 1, B → 2, C →
3, . . . , Z → 26. However, this can be easily cracked if it falls into the hands of an adversary. To add a level
of security, the messenger and receiver both agree upon a predetermined invertible matrix, say A ∈ Rn×n .
Thereafter, the sender splits up the message (without spaces!) into chunks of n consecutive letters in the
message leading to vectors in Rn . If the total number of letters is not a multiple of n, the remaining spaces
are filled by a place-holder number such as 27. Thereafter, the matrix acts on each of the vectors and yields
transformed vectors as the output, which are then sequentially transmitted to the receiver. The receiver then
decodes this message by passing each of these transformed vectors it received through A−1 . Then the message
is read by mapping the numbers to their corresponding letters.
 
1 −1
Example: Suppose the message is ‘I SEE’ and the matrix is A = ∈ R2×2 (with n = 2). Then the
2 1
sender, using an encoder machine, first converts the message into {9, 19, 5, 5}, splits it up into two vectors v1 =
[9 19]T , v2 = [5 5]T , operates on v1 and v2 using A to get y1 = Av1 = [−10 37]T and y2 = Av2 = [0 15]T , and
then transmits y1 and y2 sequentially to the receiver. The receiver then recovers v1 = A−1 y1 and v2 = A−1 y2
using a decoder, which then reads the message by replacing the numbers with suitable letters.
Suppose it is known that an adversary uses this form of coding with n = 3 but the matrix is unknown to you.
However, your spies have managed to steal a model of this encoder machine, used by the sender, and you can
now experiment with this machine (without ripping it open!) to find out what this matrix is. You decide to
send the message ‘LET ME TRY’ (without spaces, of course, and using 27 as the place-holder), and obtain a
series of numbers at the output.
(a) Determine the matrix A ∈ R3×3 if the encoder gives an output sequence {17, −8, 37, 18,
− 7, 38, 43, −9, 70} for the chosen message. Give an example of a message (not necessarily meaningful words),
using which you will certainly not be able to determine the matrix A.
(b) Suppose your friend playfully decides to send you a message using this encoder, which you read as the
sequence of numbers given by {21, −6, 36, 27, 10, 39, 23, 4, 28, 19, 0, 20, 19, 7,
24, 20, 1, 21} at the output of the encoder. What is your friend trying to tell you?

We [he and Halmos] share a philosophy about linear algebra: we think basis-free, we write basis-free, but when
the chips are down we close the office door and compute with matrices like fury.
–‘Paul Halmos: Celebrating 50 Years of Mathematics’, Irving Kaplansky