Session 2

CONCOURS D’ADMISSION 2017 FILIÈRE UNIVERSITAIRE INTERNATIONALE

MATHEMATICS
(Duration : 2 hours)

The three parts (I, II and III) are independant.

The use of computing devices is not allowed

⋆⋆ ⋆
Part I

We consider the matrix  

−1 2
A= .
−3 4

Question I.1. Determine the eigenvalues of the matrix A.

Question I.2. Determine two eigenvectors respectively associated to the two eigenvalues of A.

From now on, we consider two real numbers u0 and u1 and we define by induction the sequence
(un )n≥0 thanks to the following relations, valid for any n ≥ 1 :
u2n = 2u2n−1 − u2n−2 and u2n+1 = 3u2n − 2u2n−1 .

 
u2n
Question I.3. For n ≥ 0, we let Xn = . Express Xn+1 in terms of Xn and A.
u2n+1

Question I.4. In this question, we assume that u0 = u1 = 1. Compute the vectors Xn associated to
this choice of u0 and u1 .

Question I.5. Give a necessary and sufficient condition on the pair (u0 , u1 ) for the sequence (Xn )n≥0
to be bounded.

1
 Part II

Let f be a continuous function from [0, +∞) into R, strictly decreasing and integrable onR [0, +∞).
x
We recall that saying that the function f is integrable on [0, +∞) means that the quantity 0 |f (t)|dt
has a limit when x tends to ∞.

Question II.1. Show that the limit

lim f (x)
x→+∞
exists and is equal to zero.

Question II.2. Show that for all x ≥ 0 one has f (x) > 0.

Question II.3. Show that for all integers k ≥ 1 one has

Z k+1 Z k
f (t)dt < f (k) < f (t)dt.
k k−1

Question II.4. Deduce from the previous questions that the series, the general term of which is f (k),
converges and that one has
Z +∞ +∞
X Z +∞
∀n ≥ 1 : f (t)dt < f (k) < f (t)dt.
n k=n n−1

Question II.5. For α ∈ R and t ∈ [0, +∞) let fα (t) = 1/(t + 1)α .

II.5.a Determine the values of α for which the functionfα is integrable over [0, +∞).

II.5.b We assume that the function fα is integrable over [0, +∞). Does the following limit exist, and
if it exists, what is its value ?
+∞
X
! Z
+∞ −1
lim fα (k) fα (t)dt .
n→+∞ n
k=n

Question II.6. Does the following limit exist, and if it exists, what is its value ?
+∞
X
! Z
+∞ −1
lim exp(−k) exp(−t)dt .
n→+∞ n
k=n

Part III

We consider two polynomials P and Q with complex coefficients.

Question III.1. We assume that the polynomials P + Q and P − Q are constant. Show that the
polynomials P and Q are constant.

Question III.2. We assume that the polynomial P 2 − Q2 is a non zero constant. Show that the po-
lynomials P and Q are constant.

Question III.3. Let n ≥ 2 be an integer. We assume that the polynomial P n − Qn is a non zero
constant. Show that the polynomials P and Q are constant. Hint : one can use the factorisation of the
polynomial X n − 1 in C[X], in terms of the n-th roots of unity.

2
CONCOURS D’ADMISSION 2018 Deuxième session,

FILIÈRE UNIVERSITAIRE INTERNATIONALE

MATHEMATICS
(Duration : 2 hours)

The three parts are independent.

The use of any computing device is forbidden during this session

?? ?

Part I

In this part, n denotes a positive integer; we denote by Mn (C) the vector space of
the n × n matrices with complex entries. We consider
Pp a polynomial P of positive degree
p with complex coefficients. We write P (X) = k=0 ak X k . We are interested in matrices
A in Mn (C) such that
p
X
P (A) = 0, where P (A) = ak Ak with A0 = Idn . (1)
k=0

In this case, we say that A is a solution of (1).

I.1 Let λ be an eigenvalue of a solution A of (1). Show that one has P (λ) = 0.

I.2 Let A be a solution of (1) and Q an invertible matrix in Mn (C). Show that the
matrix Q−1 AQ is also a solution of (1).

In what follows, we consider the case n = p = 2 et P (X) = X 2 − 1; we are going to

characterise the matrices A in M2 (C) which are solution of

A2 − Id2 = 0. (2)

I.3 We first characterise the soultions of (2) which are upper triangular. Find all the
triples of complex numbers (λ, µ, α) such that the matrix
 
λ α
A=
0 µ
is a solution of (2).

I.4 Show that

 any matrix
 in M2 (C) which admits the two eigenvalues 1 et −1 is similar
1 0
to the matrix .
0 −1
I.5 Show that there exist three matrices A1 , A2 and A3 such that any solution of (2) is
either A1 or A2 , or is similar to A3 .

Part II

We consider the map f from (0, π/2) to (0, +∞) defined by

f (x) = tan x.

II.1 Show that this map is a bijection; in other words, show that this map is one-to-one
and onto.

We denote by g the map from (0, +∞) to (0, π/2) which is the reciprocal function of
f . In other words, for x in (0, π/2) and u in (0, +∞), the relations f (x) = u and g(u) = x
are equivalent.

II.2 Let u be in (0, +∞). What is the value of g(u) + g(1/u)?

Let h be the map from (0, π) into R which maps x to

 
1 x
h(x) = π − 2g √ tan .
3 2

II.3 Show that the map h is well defined and determine the set

{h(x) : x ∈ (0, π)}.

II.4 For x ∈ (0, π), compute h(h(x)).

Part III

In this part we consider the sequence of real numbers (un )n≥0 inductively defined by

1
u0 > 0 and, for any n ≥ 1 : un = un−1 + . (3)
un−1

III.1 Show that the sequence (un )n≥0 is increasing and tends to +∞ .

III.2 For n ≥ 1, express u2n − u2n−1 in terms of un−1 .

III.3 Show that for any integer n ≥ 1, one has u2n ≥ 2n.
III.4 In this question, we prove an inequality which will be used in the sequel. Show
Rk
that, for any integer k ≥ 2, one has k1 ≤ k−1 dtt . Deduce from that relation that, for any
N ≥ 1, one has
N
X 1
≤ 1 + ln N. (4)
n=1
n


III.5 Show that the sequence u2n /n n≥1 converges.


III.6 Show that there exists a real number α such that the sequence (u2n − αn)/ ln n n≥2
√
converges. Find a simple function which is equivalent to (un − αn) when n tends to in-
finity.
 CONCOURS D’ADMISSION 2019

FILIÈRE UNIVERSITAIRE INTERNATIONALE : Session d’automne

MATHEMATICS

(Duration : 2 hours)

The test is made of three independent exercises, which may be treated in any order. The use of
computing devices is not allowed.

‹‹‹

Exercise 1 : Irrationality of π
The aim of this exercise is to show that π is an irrational number. We assume by way of contradiction
a
that there exist positive integers a and b such that π “ .
b
For n P N we let Pn pXq be the polynomial
1 n
Pn pXq “ X pa ´ bXqn .
n!
.
1. Compute the coefficients of Pn .
pkq pkq
2. Show that for every k P N, Pn p0q P Z (where Pn denotes the k th derivative of Pn ).
pkq ` ˘
3. By using a symmetry of Pn , show that for every k P N, Pn ab P Z.

Let now In be defined by żπ

In “ Pn ptq sinptqdt.
0

4. Prove that In tends to 0 when n Ñ 8.

5. By using integration by parts, prove that for every n, In is an integer.
6. Conclude that π is irrational.
Exercise 2 : A calculation of the Gauss integral.
We let I be the Gauss integral ż8
2
I“ e´t dt,
0
and for x ě 0 we define ż `8
e´xt
f pxq “ ? dt.
0 tp1 ` tq

1. Show that f is a well-defined continuous function on r0, `8q.

2. Determine f p0q and lim f pxq.
xÑ`8
3. Show that f is differentiable on s0, `8q and that it satisfies the differential equation

2I
f 1 pxq ´ f pxq “ ´ ? .
x

4. Determine all the solutions of this differential equation on p0, 8q.

5. Deduce from the above the value of I.

Exercise 3 : Square roots of matrices.

In this exercise we study the existence of square roots in the space M2 pRq of 2-by-2 matrices with
real coefficients, that is for a given M P M2 pRq, we look for a matrix A with real coefficients such that

A2 “ M.

0. Prove that if λ is an eigenvalue of A, then λ2 is an eigenvalue of A2 .

1. Assume that M is diagonalizable with nonnegative eigenvalues. Show that M admits a square root
in M2 pRq.
2. Assume that M has two non-zero eigenvalues of opposite sign. Does M admit a square root in
M2 pRq ?
3. Show that ´I2 admits a square root in M2 pRq (where I2 denotes the identity matrix).
4. We now assume that M is nilpotent and M ‰ p 00 00 q.
(a) Prove that M 2 “ 0.
(b) Does M admit a square root in M2 pRq ?
5. We now assume that M admits a real non-zero eigenvalue but is not diagonalizable.
(a) Prove that there exists λ P R and a non-zero nilpotent matrix N such that M “ λI ` N .
(b) Assuming λ ą 0, prove that M admits a square root in M2 pRq.
(c) Assuming λ ă 0, prove that M does not have a square root in M2 pRq.

‹‹‹

2
 English version

CONCOURS D’ADMISSION 2020

FILIERE UNIVERSITAIRE INTERNATIONALE (FUI)

MATHEMATICS

Duration: 3 hours

The use of electronic calculators is forbidden.

Problem 1. Convergence of sequences.
Given a real sequence pun qn•1 , for every n • 1 we set

1 ÿ
n
Sn “ uk .
n k“1

We say that the sequence pun qn•1 is C-convergent if pSn qn•1 converges to some limit ` P R.
This problem deals with several aspects of the notion of C-convergence. The six numbered
questions are independent.

1. Prove carefully that if pun qn•0 converges, then it is also C-convergent, and that with the
above notation, lim un “ lim Sn .
nÑ8 nÑ8
un`1
2. (a) Let pun qn•1 be a sequence with positive values, such that converges to ` ° 0 as
un
1
n Ñ 8. Show that limnÑ8 pun q n “ `.
(b) Application: prove that pn!q n „ ne´1 as n Ñ 8.
1

3. Does C-convergence imply convergence?

4. In this question the following result is proved: if pSn qn•1 converges to ` and if furthermore
ˆ ˙
1
|un`1 ´ un | “ o ,
n
then pun q converges to `.

(a) Prove that for every n • 1 the following holds (here we put u0 “ 0)
ÿ
n ÿ
n
uk “ ´ kpuk ´ uk´1 q ` pn ` 1qun .
k“1 k“1

(b) Conclude.

5. Show that for every ` P r´1, 1s there exists a sequence pun qn•1 with values in t´1, 1u such
that pSn q converges to `. (Hint: you may use the law of large numbers.)

6. Let pxn qn•1 be the sequence defined by

$
&x “ 1;
1
%@n • 1, xn`1 “ sinpxn q.

(a) Show that pxn q is non-increasing and deduce that pxn q tends to 0.
` ˘
(b) Show that there exists a unique real number ↵ such that the sequence x↵n`1 ´ xn
↵

converges to some non-zero limit `. Give the values of ↵ and `.

c
3
(c) Deduce that xn „ as n Ñ 8.
n

1
Problem 2. A di↵erential equation.
We consider the di↵erential equation

ty 2 ` ty 1 ´ y “ 0 (E)

where y : t ބ yptq denotes a C 2 function on p0, `8q.

1. Find a polynomial solution of (E).

2. We define F on p0, `8q by

ªt
e´s
F ptq “ ds.
1 s2
(a) Study the variations of F as well as the existence of finite or infinite limits at the
boundaries of its domain.
(b) Show that the function g defined on p0, `8q by the formula

e´s 1 1
gpsq “ 2
´ 2`
s s s
admits a continuous extension at 0.
(c) Prove the existence of a constant c such that
1
F ptq “ ´ ´ ln t ` c ` op1q as t Ñ 0` .
t
1
3. We put zptq “ yptqt , where y is a solution of (E). Show that z is the solution of a linear
di↵erential equation of degree 1 and use this to deduce all the solutions of (E).

4. Show that all the solutions of (E) have a continuous extension at 0.

5. Determine the set of solutions of (E) admitting a C 1 extension at 0.

Problem 3. Linear algebra.

Let Mk,` pRq be the vector space of real matrices with k rows and ` columns and Mn pRq be
the space of square matrices of size n. In this problem we study the subspaces of Mn pRq made
of matrices of rank bounded by some r § n.
In the problem we identify a matrix M P Mn pRq and the associated linear map X ބ M X in
Mn,1 pRq » Rn . The transpose of a matrix M is denoted by M | . The identity matrix of size n
is denoted by In and Jr P Mn pRq is the following block-matrix:
˜ ¸
Ir 0
.
0 0

Recall that GLn pRq is the group of invertible n-by-n matrices.

2
1. Let M P Mn pRq. Show that the following two conditions are equivalent:

(i) M has rank r.

(ii) There exists P and Q in GLn pRq such that M “ P Jr Q.

2. Given M P Mn pRq show that M “ 0 if and only if M | M “ 0.

˜ ¸
A B
3. Let M P Mn pRq expressed in block form as , with A P GLr pRq, B P Mr,n´r pRq,
C D
C P Mn´r,r pRq and D P Mn´r pRq.

(a) Show that rankpM q • r.

˜ ¸
X
(b) Let Z P Mn,1 pRq be a vector expressed in block form as , with X P Mr,1 pRq
Y
and Y P Mn´r,1 pRq. Show that the condition Z P kerpM q is equivalent to a condition
involving Y only.
(c) Deduce that rankpM q “ r if and only if D ´ CA´1 B “ 0.

4. Let V be a vector subspace of Mn pRq such that for every M P V , rankpM q § r. We first
assume that Jr P V . Define
#˜ ¸ +
0 B
W “ , B P Mr,n´r pRq, D P Mn´r pRq .
B| D
˜ ¸
0 B
(a) Show that V X W “ t0u (Hint: given M “ P V , by using the fact that
B| D
Jr P V , show that D “ B | B “ 0.)
(b) Deduce that dimpV q § nr.

5. Let now V be an arbitrary subspace of Mn pRq such that for every M P V , rankpM q § r.
Show that dimpV q § nr.

6. Prove that this inequality is optimal.

7. Let V be a subspace such that for every M P V , rankpM q § r. We further assume that there
exists M0 P V such that rankpM0 q “ r. Let

⌦ “ tM P V, rankpM q “ ru .

Show that ⌦ is an open and dense subset of V :

(a) first assuming that M0 “ Jr ;

(b) then in the general case.

(Hint: to show the density, given M P V you may consider the line joining M and M0 . )