Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be a continuous application on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $\mathscr{S}$ be the space of solutions in $\mathscr{C}^1(\mathbb{R}, \mathbb{C}^n)$ of (2). Let $(Y_1, Y_2, \ldots, Y_n)$ be a basis of $\mathscr{S}$. For $t \in \mathbb{R}$, let $M(t)$ be the matrix whose columns are $Y_1(t), \ldots, Y_n(t)$, and let $M(t) = Q(t)\exp(tB)$ be the normal form. We admit that there exist two matrices $D$ and $N$ of $\mathscr{M}_n(\mathbb{C})$ such that $D$ is diagonalizable, $N$ is nilpotent and $B = D + N$ and $DN = ND$. There exists a matrix $P \in \mathrm{GL}_n(\mathbb{C})$ and a diagonal matrix $\Delta$ such that $D = P\Delta P^{-1}$.
For $t \in \mathbb{R}$, we denote by $Z_1(t), Z_2(t), \ldots, Z_n(t) \in \mathbb{C}^n$ the columns of the matrix $M(t)P$. Show that $(Z_1, Z_2, \ldots, Z_n)$ is a basis of the space $\mathscr{S}$.

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be a continuous application on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $\mathscr{S}$ be the space of solutions in $\mathscr{C}^1(\mathbb{R}, \mathbb{C}^n)$ of (2). Let $(Y_1, Y_2, \ldots, Y_n)$ be a basis of $\mathscr{S}$. For $t \in \mathbb{R}$, let $M(t)$ be the matrix whose columns are $Y_1(t), \ldots, Y_n(t)$, and let $M(t) = Q(t)\exp(tB)$ be the normal form. We admit that $B = D + N$ where $D$ is diagonalizable, $N$ is nilpotent, and $DN = ND$. There exists $P \in \mathrm{GL}_n(\mathbb{C})$ and a diagonal matrix $\Delta$ such that $D = P\Delta P^{-1}$. Let $Z_1(t), Z_2(t), \ldots, Z_n(t)$ be the columns of $M(t)P$.
Let $\lambda_1, \ldots, \lambda_n$ be the complex numbers such that $\Delta = \operatorname{Diag}(\lambda_1, \lambda_2, \ldots, \lambda_n)$. For all $0 \leqslant i \leqslant n-1$, $1 \leqslant k \leqslant n$ and $t \in \mathbb{R}$, we denote by $R_{i,k}(t)$ the $k$-th column of the matrix $\frac{1}{i!} Q(t) N^i P$. Show that for all $k \in \{1, 2, \ldots, n\}$, we have $$Z_k(t) = e^{\lambda_k t} \left(\sum_{i=0}^{n-1} t^i R_{i,k}(t)\right)$$ and verify that the applications $R_{i,k}$ are continuous on $\mathbb{R}$ and $T$-periodic.

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be a continuous application on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $\mathscr{S}$ be the space of solutions in $\mathscr{C}^1(\mathbb{R}, \mathbb{C}^n)$ of (2). Using the notation and results of question 17b (in particular, $Z_k(t) = e^{\lambda_k t}\left(\sum_{i=0}^{n-1} t^i R_{i,k}(t)\right)$ where $R_{i,k}$ are continuous and $T$-periodic), deduce that if the real parts of the $\lambda_i$ for $1 \leqslant i \leqslant n$ are strictly negative and if $Y$ is any solution of (2), then $$\lim_{t \rightarrow +\infty} Y(t) = 0$$

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be an application continuous on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $\mathscr{S}$ be the space of solutions in $\mathscr{C}^1(\mathbb{R}, \mathbb{C}^n)$ of (2). Let $(Y_1, Y_2, \ldots, Y_n)$ be a basis of $\mathscr{S}$. For $t \in \mathbb{R}$, let $M(t)$ be the matrix whose columns are $Y_1(t), \ldots, Y_n(t)$, and $M(t) = Q(t)\exp(tB)$ for some $B \in \mathscr{M}_n(\mathbb{C})$ and $T$-periodic $Q$. We admit that there exist two matrices $D$ and $N$ of $\mathscr{M}_n(\mathbb{C})$ such that $D$ is diagonalizable, $N$ is nilpotent and $B = D + N$ and $DN = ND$. There exists a matrix $P \in \mathrm{GL}_n(\mathbb{C})$ and a diagonal matrix $\Delta$ such that $D = P\Delta P^{-1}$.
For $t \in \mathbb{R}$, we denote by $Z_1(t), Z_2(t), \ldots, Z_n(t) \in \mathbb{C}^n$ the columns of the matrix $M(t)P$. Show that $(Z_1, Z_2, \ldots, Z_n)$ is a basis of the space $\mathscr{S}$.

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be a continuous application on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $B \in \mathscr{M}_n(\mathbb{C})$ be the matrix from the normal form $M(t) = Q(t)\exp(tB)$.
Show that if $B$ has an eigenvalue of the form $\lambda = i\frac{2k\pi}{mT}$ with $k \in \mathbb{Z}$ and $m \in \mathbb{N}^*$, then (2) has a non-zero $mT$-periodic solution.

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be a continuous application on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $B \in \mathscr{M}_n(\mathbb{C})$ be the matrix from the normal form $M(t) = Q(t)\exp(tB)$.
Suppose that there exists $m \in \mathbb{N}^*$ such that (2) has a non-zero $mT$-periodic solution. Show that $\exp(TB)$ has an eigenvalue that is an $m$-th root of unity.

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be a continuous application on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ In this question, we suppose that (2) has a $T'$-periodic solution $X$ with $T' \notin \mathbb{Q} T$.
Show that for all $t \in \mathbb{R}$ and $u \in \mathbb{R}$, we have $$A(u) X(t) = A(t) X(t).$$ One may use without proof the fact that if $G$ is a subgroup of $(\mathbb{R}, +)$ which is not of the form $\mathbb{Z}a$ for $a \in \mathbb{R}$, then $G$ is dense in $\mathbb{R}$.

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be a continuous application on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $B \in \mathscr{M}_n(\mathbb{C})$ be the matrix from the normal form $M(t) = Q(t)\exp(tB)$.
We assume that there does not exist a vector subspace $V \subset \mathbb{C}^n$, different from $\{0\}$ and $\mathbb{C}^n$, such that, for all $t \in \mathbb{R}$, $V$ is stable under $A(t)$. Give a necessary and sufficient condition on $A$ and on $B$ for (2) to have at least one non-zero periodic solution.

Let $T > 0$ be a real number and $n \in \mathbb{N}^*$ be a natural integer. Let $A : \mathbb{R} \rightarrow \mathscr{M}_n(\mathbb{C})$ be an application continuous on $\mathbb{R}$ and $T$-periodic. We consider the differential system $$X'(t) = A(t) X(t) \tag{2}$$ Let $B \in \mathscr{M}_n(\mathbb{C})$ be the matrix such that $M(t+T) = M(t)\exp(TB)$ for all $t$.
We assume in this question that there does not exist a vector subspace $V \subset \mathbb{C}^n$, different from $\{0\}$ and $\mathbb{C}^n$, such that, for all $t \in \mathbb{R}$, $V$ is invariant under $A(t)$. Give a necessary and sufficient condition on $A$ and on $B$ for (2) to have at least one non-zero periodic solution.

Solve the differential system $$\left\{ \begin{array}{l} x'(t) = x(t) - \cos(t) y(t) \\ y'(t) = \cos(t) x(t) + y(t) \end{array} \right.$$ and determine its normal form (see question 16d).

We denote by $\mathcal{E}$ the set of functions $f : \mathbb{C} \rightarrow \mathbb{C}$ expandable as a power series with radius of convergence infinity. Using the functions $Q_n \in \mathcal{E}$ satisfying $Q_n(z+1) - Q_n(z) = nz^{n-1}$ for all $n \in \mathbb{N}^*$ and $z \in \mathbb{C}$, and the bound $|Q_n(z)| \leqslant a\,\mathrm{e}^{bn|z|}$ for constants $a,b \in \mathbb{R}_+^*$, deduce the existence of a solution in $\mathcal{E}$ to the equation $(E_h)$: $$\forall z \in \mathbb{C},\, f(z+1) - f(z) = h(z)$$ when $h \in \mathcal{E}$.

A function $E$ (with rational coefficients) is a power series $f(x) = \sum_{n=0}^{\infty} \frac{b_n}{n!} x^n \in \mathbf{Q}\llbracket x \rrbracket$ satisfying: (a) $f$ is a solution of a differential equation; (b) there exists a real number $C > 0$ such that $|b_n| \leq C^n$ and $\operatorname{denom}(b_0, \ldots, b_n) \leq C^n$ for all $n \geq 1$.
Recall that $\widehat{f}(x)$ denotes the Laplace transform $\widehat{f}(x) \stackrel{\text{def}}{=} \sum_{n=0}^{\infty} b_n x^n$. Which functions $E$ are such that $\widehat{f}$ is also a function $E$?

A function $E$ (with rational coefficients) is a power series $f(x) = \sum_{n=0}^{\infty} \frac{b_n}{n!} x^n \in \mathbf{Q}\llbracket x \rrbracket$ satisfying: (a) $f$ is a solution of a differential equation; (b) there exists a real number $C > 0$ such that $|b_n| \leq C^n$ and $\operatorname{denom}(b_0, \ldots, b_n) \leq C^n$ for all $n \geq 1$.
Prove that functions $E$ are closed under addition and multiplication.

A function $E$ (with rational coefficients) is a power series $f(x) = \sum_{n=0}^{\infty} \frac{b_n}{n!} x^n \in \mathbf{Q}\llbracket x \rrbracket$ satisfying: (a) $f$ is a solution of a differential equation; (b) there exists a real number $C > 0$ such that $|b_n| \leq C^n$ and $\operatorname{denom}(b_0, \ldots, b_n) \leq C^n$ for all $n \geq 1$.
Let $f$ be an exponential polynomial (i.e., $f(x) = \sum_{i=1}^{s} P_i(x) e^{c_i x}$ with $c_i \in \mathbf{Q}$ and $P_i \in \mathbf{Q}[x]$). Show that $f$ is a function $E$ such that $\widehat{f}$ is the power series expansion of a rational fraction with rational poles.

A function $E$ (with rational coefficients) is a power series $f(x) = \sum_{n=0}^{\infty} \frac{b_n}{n!} x^n \in \mathbf{Q}\llbracket x \rrbracket$ satisfying: (a) $f$ is a solution of a differential equation; (b) there exists a real number $C > 0$ such that $|b_n| \leq C^n$ and $\operatorname{denom}(b_0, \ldots, b_n) \leq C^n$ for all $n \geq 1$.
Show that if $\sum_{n=0}^{\infty} b_n x^n$ is the power series expansion of a rational fraction with rational poles, then $\sum_{n=0}^{\infty} \frac{b_n}{n!} x^n$ is a function $E$.

A function $E$ (with rational coefficients) is a power series $f(x) = \sum_{n=0}^{\infty} \frac{b_n}{n!} x^n \in \mathbf{Q}\llbracket x \rrbracket$ satisfying: (a) $f$ is a solution of a differential equation; (b) there exists a real number $C > 0$ such that $|b_n| \leq C^n$ and $\operatorname{denom}(b_0, \ldots, b_n) \leq C^n$ for all $n \geq 1$.
Show that the Bessel function $$J_0(x) \stackrel{\text{def}}{=} \sum_{n=0}^{\infty} \frac{(-1)^n}{(n!)^2} \left(\frac{x}{2}\right)^{2n}$$ is a function $E$ such that $\widehat{J}_0(x)$ satisfies the equation $(1 + x^2)\widehat{J}_0(x)^2 = 1$. Deduce that $J_0(x)$ is not an exponential polynomial.

Show that the real zeros of the Bessel function $J_0(x) \stackrel{\text{def}}{=} \sum_{n=0}^{\infty} \frac{(-1)^n}{(n!)^2} \left(\frac{x}{2}\right)^{2n}$ are simple, that is, if $J_0(\alpha) = 0$, then $J_0'(\alpha) \neq 0$.

129. The function with the rule $f(x) = \displaystyle\lim_{n \to +\infty} \left(1 - \dfrac{3x}{n}\right)^n$ is defined for every real number $x$. Which statement is correct?

• [(1)] $f''(x) + 6f'(x) + 9f(x) = 0$
• [(2)] $f''(x) + 3f'(x) + 2f(x) = 0$
• [(3)] $f''(x) - 6f'(x) + 9f(x) = 0$
• [(4)] $f''(x) - 3f'(x) + 2f(x) = 0$

10. Determine which are the values of the parameter $k \in \mathbb{R}$ for which the function $y(x) = 2e^{kx+2}$ is a solution of the differential equation $y'' - 2y' - 3y = 0$.
\footnotetext{Maximum duration of the examination: 6 hours. The use of scientific and/or graphical calculators is permitted provided they are not equipped with symbolic calculation capacity (O.M. no. 350 Art. 18 paragraph 8). The use of a bilingual dictionary (Italian–language of the country of origin) is permitted for candidates whose native language is not Italian. It is not permitted to leave the Institute before 3 hours have elapsed from the dictation of the theme.}

Let $f : [ 0,1 ] \rightarrow \mathbb { R }$ (the set of all real numbers) be a function. Suppose the function $f$ is twice differentiable, $f ( 0 ) = f ( 1 ) = 0$ and satisfies $f ^ { \prime \prime } ( x ) - 2 f ^ { \prime } ( x ) + f ( x ) \geq e ^ { x } , x \in [ 0,1 ]$.
Which of the following is true for $0 < x < 1$?
(A) $0 < f ( x ) < \infty$
(B) $- \frac { 1 } { 2 } < f ( x ) < \frac { 1 } { 2 }$
(C) $- \frac { 1 } { 4 } < f ( x ) < 1$
(D) $- \infty < f ( x ) < 0$

Let $f : [ 0,1 ] \rightarrow \mathbb { R }$ (the set of all real numbers) be a function. Suppose the function $f$ is twice differentiable, $f ( 0 ) = f ( 1 ) = 0$ and satisfies $f ^ { \prime \prime } ( x ) - 2 f ^ { \prime } ( x ) + f ( x ) \geq e ^ { x } , x \in [ 0,1 ]$.
If the function $\mathrm { e } ^ { - x } f ( x )$ assumes its minimum in the interval $[ 0,1 ]$ at $x = \frac { 1 } { 4 }$, which of the following is true?
(A) $f ^ { \prime } ( x ) < f ( x ) , \frac { 1 } { 4 } < x < \frac { 3 } { 4 }$
(B) $f ^ { \prime } ( x ) > f ( x ) , \quad 0 < x < \frac { 1 } { 4 }$
(C) $f ^ { \prime } ( x ) < f ( x ) , \quad 0 < x < \frac { 1 } { 4 }$
(D) $f ^ { \prime } ( x ) < f ( x ) , \frac { 3 } { 4 } < x < 1$

Let $f ( x )$ be a positive function such that the area bounded by $y = f ( x ) , y = 0$ from $x = 0$ to $x = a > 0$ is $e ^ { - a } + 4 a ^ { 2 } + a - 1$. Then the differential equation, whose general solution is $y = c _ { 1 } f ( x ) + c _ { 2 }$, where $c _ { 1 }$ and $c _ { 2 }$ are arbitrary constants, is
(1) $\left( 8 e ^ { x } - 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } + \frac { d y } { d x } = 0$
(2) $\left( 8 e ^ { x } - 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } - \frac { d y } { d x } = 0$
(3) $\left( 8 e ^ { x } + 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } - \frac { d y } { d x } = 0$
(4) $\left( 8 e ^ { x } + 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } + \frac { d y } { d x } = 0$

Q74. Let $\int _ { 0 } ^ { x } \sqrt { 1 - \left( y ^ { \prime } ( t ) \right) ^ { 2 } } d t = \int _ { 0 } ^ { x } y ( t ) d t , 0 \leq x \leq 3 , y \geq 0 , y ( 0 ) = 0$. Then at $x = 2 , y ^ { \prime \prime } + y + 1$ is equal to
(1) 1
(2) 2
(3) $\sqrt { 2 }$
(4) $1 / 2$

Q75. Let $f ( x )$ be a positive function such that the area bounded by $y = f ( x ) , y = 0$ from $x = 0$ to $x = a > 0$ is $e ^ { - a } + 4 a ^ { 2 } + a - 1$. Then the differential equation, whose general solution is $y = c _ { 1 } f ( x ) + c _ { 2 }$, where $c _ { 1 }$ and $c _ { 2 }$ are arbitrary constants, is
(1) $\left( 8 e ^ { x } - 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } + \frac { d y } { d x } = 0$
(2) $\left( 8 e ^ { x } - 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } - \frac { d y } { d x } = 0$
(3) $\left( 8 e ^ { x } + 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } - \frac { d y } { d x } = 0$
(4) $\left( 8 e ^ { x } + 1 \right) \frac { d ^ { 2 } y } { d x ^ { 2 } } + \frac { d y } { d x } = 0$

Problem 1
I. Find the general solution of the following differential equation:
$$\frac { d ^ { 4 } y } { d x ^ { 4 } } - 2 \cdot \frac { d ^ { 3 } y } { d x ^ { 3 } } + 2 \frac { d y } { d x } - y = 9 e ^ { - 2 x }$$
Here, $e$ denotes the base of the natural logarithm.
II. Find the value of the following integral:
$$\int _ { 0 } ^ { 1 } x ^ { m } ( \log x ) ^ { n } d x$$
Here, $m$ and $n$ are non-negative integers.
III. We define $I ( m )$ as
$$I ( m ) \equiv \int _ { 0 } ^ { 1 } x ^ { m } \arccos x \, d x$$
Here, $m$ is a non-negative integer. Use the principal values of inverse trigonometric functions.

1. Find the value of $I ( 0 )$.
2. Find the value of $I ( 1 )$.
3. Express $I ( m )$ in terms of $m$ and $I ( m - 2 )$ when $m \geq 2$.
4. Find the value of $I ( m )$.