Skip to content

Instantly share code, notes, and snippets.

@rricard
Created December 21, 2011 23:08
Show Gist options
  • Save rricard/1508125 to your computer and use it in GitHub Desktop.
Save rricard/1508125 to your computer and use it in GitHub Desktop.
Démonstrations de Mathématiques de Première Année à l'INSA de Lyon
\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[francais]{babel}
\usepackage{amsthm}
\usepackage{amsmath}
\usepackage{amssymb}
\usepackage{mathrsfs}
\usepackage{mathpazo}
\usepackage[scaled=.95]{helvet}
\usepackage{courier}
%\newcommand{\deg}{\mathrm{deg }}
\newcommand{\ch}{\mathrm{ch }}
\newcommand{\sh}{\mathrm{sh }}
\newcommand{\thy}{\mathrm{th }}
\newcommand{\argch}{\mathrm{argch }}
\newcommand{\argsh}{\mathrm{argsh }}
\newcommand{\argth}{\mathrm{argth }}
\title{Démonstrations de Mathématiques de Première Année à l'INSA de Lyon}
\author{Robin Ricard\\
Etudiant en première année,\\
Département Premier Cycle,\\
INSA de Lyon\\
\texttt{robin.ricard@insa-lyon.fr}}
\date{\today}
\begin{document}
\maketitle
\begin{it}
Merci de me signaler toute erreur ou complément éventuel à apporter à ce document.\\
Ce document est librement utilisable selon la license Creative Commons CC-BY-NC-3.0 (vous ne pouvez pas en faire un usage commercial et vous devez citer son auteur si vous l'utilisez).
\end{it}
\section{Nombres Complexes}
\subsection{Existence et forme des racines $n$-ièmes d'un nombre complexe non nul}
Soit $Z$ un nombre complexe non nul, définissons $z$ tel que $z^n=Z$. Vérifions son existence. \\
Sachant que $Z$ est de la forme :
\[Z=\rho e^{i(\theta+2k\pi)}\]
$z$ existe alors et est de la forme :
\[z=\sqrt[n]{\rho}e^{i\frac{\theta+2k\pi}{n}}\]
%Existence et forme des racines $n$-ièmes d'un nombre complexe non nul subsection (end)
\subsection{La somme des racines $n$-ièmes d'un nombre complexe non nul est nulle}
Soit $z_k$ une racine $n$-ième de $Z$. On a d'après (1.1) :
\[z_k=\sqrt[n]{\rho}e^{i\frac{\theta+2k\pi}{n}}\]
On peut dire :
\[\sum^{n-1}_{k=0}z_k=
\sum^{n-1}_{k=0}\sqrt[n]{\rho}e^{i\frac{\theta}{n}}e^{i\frac{2k\pi}{n}}=
\sqrt[n]{\rho}e^{i\frac{\theta}{n}}\sum^{n-1}_{k=0}(e^{i\frac{2\pi}{n}})^k=
\sqrt[n]{\rho}e^{i\frac{\theta}{n}}\frac{1-(e^{i\frac{2\pi}{n}})^n}{1-e^{i\frac{2\pi}{n}}}\]
\[=\sqrt[n]{\rho}e^{i\frac{\theta}{n}}\frac{1-1}{1-e^{i\frac{2\pi}{n}}}=
0\]
On a aussi un autre moyen de le montrer :
\[{z_k}^n=Z \Leftrightarrow
{z_k}^n-Z=0 \Leftrightarrow
{z_k}^n+0{z_k}^{n-1}-Z=0
\]
On a fait apparaître que la somme des racines de ce polynôme (qui n'est autre que la sommes des racines du nombre complexe) est nulle. (par l'utilisation des relations coefficients/racines d'un polynôme)
%La somme des racines $n$-ièmes d'un nombre complexe non nul est nulle subsection (end)
%Nombres Complexes section (end)
\section{Polynômes}
\subsection{Unicité du couple Quotient/Reste dans la division euclidienne}
Admettons (par l'absurde) que $(Q', R')$ soit un autre couple de solution à l'équation $A=BQ+R$ où $(Q,R)$ est un couple solution. \\
On a :
\[0=B(Q-Q')+(R-R')
\Rightarrow B(Q-Q')=R'-R\]
\[\deg(R'-R)=\deg(B)+\deg(Q-Q')\]
Or, $\deg(R'-R)$ est inférieur à $\max(\deg(R'),\deg(R))$ donc strictement inférieur à $\deg(B)$. \\
Obligatoirement, on a $Q'=Q$ (pour avoir le résultat de la soustraction au degré 0) et donc $R'=R$. \\
Le couple solution est bien unique lorsque l'on effectue la division euclidienne de deux polynômes.
%Unicité du couple Quotient/Reste dans la division euclidienne subsection (end)
%Polynômes section (end)
\section{Suites Numériques}
\subsection{Toute suite croissante et majorée est convergente}
On a :
\[\exists N | \forall n \geqslant N, U_{n+1} \geqslant U_n\]
\[\exists M | \forall n, |U_n| \leqslant M\]
Soit $Z=\{U_n, n\in \mathbb{N}\}$ non vide ($U_0 \in Z$) et majoré donc $Z$ admet la borne supérieure $B$ définie par :
\[\forall n, U_n\leqslant B\]
\[\forall \epsilon > 0, \exists N_\epsilon | B-\epsilon<U_{N_\epsilon}<B\]
A partir d'un certain rang, $\max(N_\epsilon,N)$, on a :
\[B>U_{n+1}>U_n\]
\[B-\epsilon<U_{N_\epsilon}<B\]
\[\Rightarrow \forall \epsilon, \exists N_\epsilon, n>N_\epsilon\]
\[\Rightarrow |U_n-B|<\epsilon\]
$\Rightarrow$ $(U_n)$ converge vers $B$
%Toute suite croissante et majorée est convergente subsection (end)
\subsection{Deux suites adjacentes sont convergentes et ont la même limite}
Supposons (par l'absurde) que $\forall N | U_N<V_N$ \\
Soit $(W_n)=(U_n-V_n)$.
Par hypothèse, on a : $(U_n)$ décroissante et $(V_n)$ croissante, $(W_n)$ tend vers 0. \\
De plus $(W_n)$ décroît. \\
Mais on voit aussi que $\exists N | W_n<0$. Ce qui est contradictoire. \\
Donc $\forall N | U_N>V_N$ et donc $(U_n)$ est minorée par $V_0$. \\
$(U_n)$ est minorée et décroît donc converge. \\
De la même manière $(V_n)$ converge. \\
Comme $(Un-Vn)$ tend à être nul, les deux suites ont la même limite.
%Deux suites adjacentes sont convergentes et ont la même limite subsection (end)
\subsection{Si la fonction $f$ est croissante, une suite définie par la relation $U_{n+1}=f(U_n)$ est monotone}
$f$ est croissante donc on a :
\[x<y \Rightarrow f(x)\leqslant f(y)\]
$U_0$ est fixé, on connaît donc $U_1$. Par disjonction des cas, on a :
\begin{itemize}
\item $U_0>U_1$ \\
$f(U_0) \geqslant f(U_1) \Leftrightarrow U_1 \geqslant U_2$ car $f$ croissante \\
Par récurrence on montre que : $\forall n \in \mathbb{N} | U_{n+1} \leqslant U_n$ \\
$(U_n)$ décroît et est monotone
\item $U_0<U_1$ \\
$f(U_0) \leqslant f(U_1) \Leftrightarrow U_1 \leqslant U_2$ car $f$ croissante \\
Par récurrence on montre que : $\forall n \in \mathbb{N} | U_{n+1} \geqslant U_n$ \\
$(U_n)$ croît et est monotone
\item $U_0=U_1$ \\
On a alors $\forall n \in \mathbb{N} | U_n=U_0$ \\
$(U_n)$ constante et est monotone
\end{itemize}
$(U_n)$ est monotone
%Si la fonction $f$ est croissante, une suite définie par la relation $U_{n+1}=f(U_n)$ est monotone subsection (end)
%Suites Numériques section (end)
\section{Fonctions Usuelles}
\subsection{Dérivabilité et expression de la dérivée des fonctions $\arcsin$, $\arccos$, $\arctan$, $\argsh$, $\argch$, $\argth$}
\subsubsection{$\arcsin$}
On utilise la formule :
\[(f^{-1})'(y_0)=\frac{1}{f'(f^{-1}(y_0))}\]
On rapelle aussi que :
\[\cos^2 x +\sin^2 x=1\]
Soit $x_0 \in [0;\pi]$ et $y_0=\sin x_0$
\[\arcsin'(y_0)=\frac{1}{\cos(\arcsin y_0)}=\frac{1}{\cos x_0}=\frac{1}{\sqrt{1-\sin^2 x_0}}=\frac{1}{\sqrt{1-{y_0}^2}}\]
%$\arcsin$ subsubsection (end)
\subsubsection{$\arccos$}
Soit $x_0 \in [0;\pi]$ et $y_0=\cos x_0$
\[\arccos'(y_0)=-\frac{1}{\sin(\arccos y_0)}=-\frac{1}{\sin x_0}=-\frac{1}{\sqrt{1-\cos^2 x_0}}=-\frac{1}{\sqrt{1-{y_0}^2}}\]
%$\arcsin$ subsubsection (end)
\subsubsection{$\arctan$}
On rapelle que :
\[\tan'x=1+\tan^2x\]
Donc, avec la formule :
\[\arctan'x=\frac{1}{1+\tan^2(\arctan x)}=\frac{1}{1+x^2}\]
%$\arctan$ subsubsection (end)
\subsubsection{$\argsh$}
On rapelle que :
\[\ch^2 x -\sh^2 x=1\]
Soit $x_0 \in [0;\pi]$ et $y_0=\sh x_0$
\[\argsh'(y_0)=\frac{1}{\ch(\argsh y_0)}=\frac{1}{\ch x_0}=\frac{1}{\sqrt{\sh^2 x_0+1}}=\frac{1}{\sqrt{{y_0}^2+1}}\]
%$\argsh$ subsubsection (end)
\subsubsection{$\argch$}
Soit $x_0 \in [0;\pi]$ et $y_0=\ch x_0$
\[\argch'(y_0)=\frac{1}{\sh(\argch y_0)}=\frac{1}{\sh x_0}=\frac{1}{\sqrt{\ch^2 x_0-1}}=\frac{1}{\sqrt{{y_0}^2-1}}\]
%$\argch$ subsubsection (end)
\subsubsection{$\argth$}
On rapelle que :
\[\thy'x=1-\thy^2x\]
Donc, avec la formule :
\[\argth'x=\frac{1}{1-\thy^2(\argth x)}=\frac{1}{1-x^2}\]
%$\argth$ subsubsection (end)
%Dérivabilité et expression de la dérivée des fonctions $\arcsin$, $\arccos$, $\arctan$, $\argsh$, $\argch$, $\argth$ subsection (end)
%Fonctions Usuelles section (end)
\section{Comparaison de Fonctions}
\textit{Ceci n'est qu'interprétation de ce que l'on peut nous demander. L'énoncé étant ambigü...}
\subsection{Logarithme et équivalence (compatibilité avec le logarithme)}
Soient $f$ et $g$ définies et positives en $a$ telles que $f\underset{a}{\sim}g$
\subsubsection{si $\lim\limits_{x\to a} f\neq 1$}
\[\ln{g}=\ln(\frac{g}{f}f)=\ln{\frac{g}{f}}+\ln{f}\]
Et :
\[\lim\limits_{x\to a}\ln{\frac{g}{f}}=0\]
On a :
\[\frac{\ln{g}}{\ln{f}}=\frac{\ln{\frac{g}{f}}}{\ln{f}}+1\]
Logiquement :
\[\lim\limits_{x\to a}\frac{\ln{g}}{\ln{f}}=1
\Rightarrow \ln{f} \underset{a}{\sim} \ln{g}\]
%si $\lim\limits_{x\to a} f\neq 0$ subsubsection (end)
\subsubsection{si $\lim\limits_{x\to a} f=1$}
On sait que :
\[\ln{x} \underset{1}{\sim} x-1\]
A l'aide de la compatibilité des équivalences avec la composition à droite, on a :
\[\ln(g(x)) \underset{a}{\sim} g(x)-1\]
De plus, par hypothèse :
\[\lim\limits_{x\to a}f=\lim\limits_{x\to a}g=1\]
Donc :
\[\ln(f(x)) \underset{a}{\sim} f(x)-1\]
%si $\lim\limits_{x\to a} f=0$ subsubsection (end)
%Logarithme et équivalence (compatibilité avec le logarithme) subsection (end)
\subsection{Compatibilité avec les puissances (non requise)}
Soient $f$ et $g$ définies et positives en $a$ telles que $f\underset{a}{\sim}g$
\subsubsection{$\forall \alpha \in \mathbb{R}^*$}
On a :
\[\lim\limits_{x\to a}\frac{g}{f}=1\]
Et :
\[\frac{g^\alpha}{f^\alpha}=(\frac{g}{f})^\alpha\]
Si et seulement si :
\[\lim\limits_{x\to a} \frac{g^\alpha}{f^\alpha} = 1
\Leftrightarrow f^\alpha \underset{a}{\sim} g^\alpha\]
%$\forall \alpha \in \mathbb{R}^*$ subsection (end)
\subsubsection{$\forall \alpha > 0$}
Avec $\epsilon_k \to 0$
\[f\underset{a}{=}o(g)
\Leftrightarrow f\underset{a}{=}g\epsilon_1(x)
\Leftrightarrow f^\alpha\underset{a}{=}g^\alpha\epsilon_2(x)
\Leftrightarrow f^\alpha\underset{a}{=}o(g^\alpha)\]
%$\forall \alpha \in > 0$ subsection (end)
\subsubsection{$\forall \alpha < 0$}
Avec $\epsilon_k \to 0$ et $\alpha=-\beta$
\[f\underset{a}{=}o(g)
\Leftrightarrow f\underset{a}{=}g\epsilon_1(x)
\Leftrightarrow f^\beta\underset{a}{=}g^\beta\epsilon_2(x)
\Leftrightarrow \frac{1}{f^\alpha}\underset{a}{=}\frac{1}{g^\alpha}\epsilon_2(x)
\Leftrightarrow g^\alpha\underset{a}{=}f^\alpha\epsilon_2(x)
\Leftrightarrow g^\alpha\underset{a}{=}o(f^\alpha)\]
%$\forall \alpha \in < 0$ subsection (end)
%Compatibilité avec les puissances (non requise) subsection (end)
\subsection{Compatibilité avec exponentielle}
Soit $f \underset{a}{\sim} g$
\[\lim\limits_{x \to a} f(x)-g(x)=0\]
\[\Leftrightarrow \lim\limits_{x \to a} e^{f(x)-g(x)}=1\]
\[\Leftrightarrow \lim\limits_{x \to a} e^{f(x)}(e^{g(x)})^{-1}=1\]
\[\Leftrightarrow \lim\limits_{x \to a} \frac{e^{f(x)}}{e^{g(x)}}=1\]
\[\Leftrightarrow e^{f(x)} \underset{a}{\sim} e^{g(x)} \]
%Compatibilité avec exponentielle subsection (end)
%Comparaison de Fonctions section (end)
\section{Dérivation}
\subsection{Théorème de Rolle}
On a :
\begin{itemize}
\item $f$ continue sur $[a,b]$
\item $f$ dérivable sur $]a,b[$
\item $f(a)=f(b)$
\end{itemize}
Puisque $f$ est continue, alors l'image de $[a,b]$ par $f$ est un intervalle fermé $[m,M]$
\begin{itemize}
\item $m=M$ \\
alors $f$ est constante sur $[a,b]$ donc $f'$ est nulle sur $]a,b[$
\[\forall c \in ]a,b[ | f'(c)=0\]
\item $m<M$
\[\exists c_M \in ]a,b[ | f(c_M)=M\]
On a :
\[\forall x \in [a,b], f(x)\leqslant f(c_M)\]
Or, $f'(c_M)$ existe ($f'$ est bien dérivable dans cet intervalle), donc :
\[\lim\limits_{\substack{x \to c_M \\ x>c_M}} \frac{f(x)-f(c_M)}{x-c_M}=f'(c_M)\]
Soit $x<c_M$, $f(x)-f(c_M)\leqslant 0$ et $x-c_M \leqslant 0$, donc :
\[\lim\limits_{\substack{x \to c_M \\ x>c_M}} \frac{f(x)-f(c_M)}{x-c_M}=f'(c_M) \geqslant 0\]
De même $x>c_M$ implique que :
\[f'(c_M) \leqslant 0\]
Comme on peut avoir n'importe quelle valeur de $x$, on doit forcément prendre :
\[f'(c_M)=0\]
\end{itemize}
%Théorème de Rolle subsection (end)
\subsection{Théorème des accroissements finis}
On a :
\begin{itemize}
\item $f$ continue sur $[a,b]$
\item $f$ dérivable sur $]a,b[$
\end{itemize}
Soit :
\[g: x \mapsto f(x)-\frac{f(b)-f(a)}{b-a}(x-a)\]
$g$ est continue et dérivable comme $f$ \\
De plus $g(b)=f(a)$ et $g(a)=f(a)$. On peut dire que : $g(b)=g(a)$ \\
On applique alors le théorème de Rolle à $g$ sur $[a,b]$.
\[\exists c \in ]a,b[ | g'(c)=0\]
Or :
\[g'(c)=f'(c)-\frac{f(b)-f(a)}{b-a}=0\]
D'où :
\[f(b)-f(a)=f'(c)(b-a)\]
%Théorème des accroissements finis subsection (end)
%Dérivation section (end)
\end{document}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment