Math worksheet 3.

2 - Linear maps
elements of correction

Ecam Engineering 2 - S4 – 2022-2023

Exercise 1

For all the following maps, show that they are linear, find their Kernel, their Image, find out
whether they are surjective/injective/an isomorphism and verify the rank nullity theorem if
appropriate.

1.
 

a : R3 → R3  1 0 0 
where A =  1 1 1
 

X 7→ AX  
1 2 −1

(linear by linearity of matrix product)

This matrix is invertible (determinant is not 0, OR make CR decomposition and obtain

3 pivots...), hence it is an isomorphism, its kernel contains only the 0R3 and its image
is R3 .

2.
 
1 −10 −24 −42
b : R4 → R3  
where B =  1 −8 −18 −32 
 
X 7→ BX  
−2 20 51 87

(linear by linearity of matrix product)

     
 1  −10 −24
Create a CR decomposition, you’ll get 3 pivots and then Col(B) = span{ 1  ,  −8  , −18}
     
     
−2 20 51
 
−2
 
−2
then its rank is 3, therefore it is a surjection. You should find kerA = span{ }.
 
−1
 
1

3.

d : R3 [X] → R3 [X]
f 7→ f ′

(linear by linearity of differentiation)

1
 Write explicitely P ∈ R3 [X] : P (x) = ax3 + bx2 , +cx + d, you can calculate the effect of
map d on P .. and clearly see that images are spanned by {1, X, X 2 } hence Im d = R2 [X]
and also Ker d = span{1}, then this map is neither injective nor surjective.

4. same map but with co-domain R2 [X]. Now the map becomes surjective.

5. same map but with domain C ∞ (R) and co-domain also C ∞ (R) surjective but not
injective.

6.

i : R3 [X] → R4 [X]
Z x
f 7→ f (t)dt
0

(linear by linearity of integration) this map gives the primitive that cancels at 0.. then it
is going to be injective but not surjective (as the images are polynomials with constant
term equal to 0)

7. same map but with domain C 0 (R) to C00 (R) (ie C 0 functions equal to 0 at 0) injective
Rx Rx
(if 0 f (t)dt = 0 g(t)dt , differentiate both sides, you get f (x) = g(x) everywhere),
but not surjective as image belongs to C 1 and there are functions that are not C 1 in
C 0 ).

8. same map but with domain C 0 (R) to C01 (R) (ie C 1 functions equal to 0 at 0) this
time, it is an isomorphism

9.

e : C 1 (R) → C 1 (R)
f 7→ f ′ + f

Ker e = span {e−t } (solving f ′ + f = 0), so it is not injective. And using variation of
parameters you can always solve f ′ + f = g for any g ∈ C 1 , therfore it is surjective.

Exercise 2

On E = R3 , we define the map f given by

∀(x, y, z) ∈ R3 f (x, y, z) = (4x + y − z, 2x + 3y − z, 2x + y + z)

2
1- Show that f ∈ L(R3 ) and make its representing matrix in the canonical basis of R3
 
4 1 −1
Mf = 2 3 −1
 
 
2 1 1

2- Find two distinct real numbers λ and µ such that f − λ idE and f − µ idE are not
automorphisms. f − λ idE not an automorphism means it is not bijective and in particular
we are looking for values such that Mf − λI is not invertible .. therfore we calculate the
determinant, (and get the characteristic equation, we are indeed looking for eigenvalues of
the matrix...) and obtain values 4 and 2.
3- Show that E = Ker(f − λ idE ) ⊕ Ker(f − µ idE ) We find basis for Ker(f − 4 idE ) =
span{(1, 1, 1)} and Ker(f − 2 idE ) = span{(1, 0, 2), (0, 1, 1)} . Actually, we can show for
instance that those 3 vectors are linearly independent (for instance, the determinant of the
passage matrix that contains them as columns is not 0)... so they make a basis of R3 and
thus the spanned spaces make a direct sum ( as the decomposition is unique).
4- Find the representing matrix of f in a basis adapted to this direct sum. Well then in the
adapted basis  
4 0 0 
0 2 0 
 
 
0 0 2

yes it is diagonalisation.

Exercise 3

Given n ∈ N and ψ a map from Rn [X] to Rn [X], defined by:

∀P ∈ Rn [X], ψ(P ) = P (X + 1) − P (X)

1- Show that ψ ∈ L(Rn [X]).

2- a) Express the degree of ψ(P ) for any P ∈ Rn [X]. Let’s find the image of the canonical
basis polynomials:
   
k k−1
X k X k
ψ(X k ) = (X + 1)k − X k =   Xj − Xk =   Xj
j=0 j j=0 j

Thus the degree of ψ(X k ) is (k − 1).

b) Find the image and kernel of ψ. you can create for instance the matrix representation
.. it will be an (n + 1) × (n + 1) upper triangular matrix with zeros on the diagonal and the

3
Pascal’s triangle coefficients above (see previous calculation with the binomial formula !). It
is a row echelon matrix with n non zero pivots.. we obtain:

Ker ψ = span {1}

Im ψ = Rn−1 [X]

3- Given P ∈ Rn [x] of degree n, show that (P, ψ(P ), ψ 2 (P ), · · · , ψ n (P )) is a basis of Rn [X].

The degree of ψ(P ) is (n-1), hence the degree of ψ k (P ) is (n − k), then we have a family of
polynomials of echeloned degree... then they are all linearly independant, and having (n + 1)
of them makes a basis of Rn [X].
4 a) Given Q in Rn−1 [X], show that there exists a unique polynomial P ∈ Rn [X] such that:

P (X + 1) − P (X) = Q(X) and P (0) = 0

From previous questions, we saw that psi was surjective on Rn−1 [X]. So existence of P
given Q is granted. Now to get uniqueness, we proceed as usual: if ψ(P1 ) = ψ(P2 ) then
(P1 − P2 ) ∈ kerψ: in other words P1 and P2 differ by a constant (from Q 2-b), but if they
are both 0 at 0.. they must be equal.
b) Find such a polynomial P for Q(X) = X(X + 1)(X + 2), and deduce a formula for

n
X
k(k + 1)(k + 2)
k=0

We are looking for a polynomial of R4 [X] (as degree of P is 3) with 0 constant term. Hence
we are solving P (X + 1) − P (X) = X(X + 1)(X + 2) with P (X) = aX 4 + bX 3 + cC 2 + dX.
We find : P = 41 X 4 + 21 X 3 − 14 X 2 − 12 X
Enventually:

n n
X X 1 1 1 1
k(k+1)(k+2) = P (k+1)−P (k) = P (n+1)−P (0) = (n+1)4 + (n+1)3 − (n+1)2 − (n+1)
4 2 4 2
k=0 k=0

Exercise 4

Denote E the vector space of continuous functions on R, and T the map on E, such that for
any f ∈ E, T (f ) = F where F is defined by:
Z x
2
∀x ∈ R, F (x) = f (t)e−t dt
0

Show that T is an injective endomorphism on E. Is it surjective? you have to prove :

4
 − linearity : yes by linearity of integration.. but you still have to write the details.

− endomorphism (image is also continuous) : indeed if f is continuous then F (x) as

defined exists and F is differentiable (fundamental theorem of calculus, its derivative
2 /2
F ′ (x) = f (x)e−x therefore continuous. (note that we already know it won’t be
surjective as there are continuous functions that are not differentiable).

To show injectivity, we are looking for the kernel of T : functions f such that T (f ) = 0C 0
(image is the 0 cosntant function), i.e. By differentiation, we have that F ′ (x) = 0 for all x :
2 /2
f (x)e−x = 0, it yields f (x) = 0 for all x ∈ R.. therefore T is injective.

The next 2 exercises are if you want to go further in your understanding... come and see me
for a correction if needed.

Exercise 5

Given a vector space E, and f and g in L(E), show that:

1- Ker(f ) ⊂ Ker(g ◦ f )
2- Im(g ◦ f ) ⊂ Im(g)
3- f (Ker(g ◦ f )) = Ker g ∩ Im(f )
classically, if you want to prove A ⊂ B, you need to prove that ∀X ∈ A, X ∈ B. (and to
prove A = B you need A ⊂ B and B ⊂ A).

Exercise 6

Given a n-dimensional vector space E, and f and g two endomorphisms on E such that:
(i) g ◦ f = 0
(ii) f + g is a bijective endomorphism (= automorphism)
Show that rank f + rank g = n