📚 الرئيسية

مسابقة دكتوراه 2022Université Ferhat Abbas - Sétif 1 — الموضوع 01

مسابقة عامة · الرياضيات

Concours national d'accès au Doctorat 2022-2023, Faculté des Sciences, Épreuve : Algèbre Linéaire

التمرين 1

Exercice 1 — Endomorphismes qui commutent : $u\circ f + v\circ g = \mathrm{id}_E$

#endomorphismes#noyaux#images#décomposition

Soient EE un espace vectoriel et u,v,f,gu,v,f,g quatre endomorphismes de EE qui deux à deux commutent, et vérifient uf+vg=idEu\circ f + v\circ g = \mathrm{id}_E.

  1. Montrer que : kerfkerg={0}\ker f \cap \ker g = \{0\} et E=Imf+ImgE = \mathrm{Im}\, f + \mathrm{Im}\, g.

  2. Montrer que : ker(fg)=kerfkerg\ker(f\circ g) = \ker f \oplus \ker g et Im(fg)=ImfImg\mathrm{Im}(f\circ g) = \mathrm{Im}\, f \cap \mathrm{Im}\, g.

  3. On suppose de plus que fg=0f\circ g = 0. Montrer que : E=kerfkerg=ImfImgE = \ker f \oplus \ker g = \mathrm{Im}\, f \oplus \mathrm{Im}\, g, kerf=Img\ker f = \mathrm{Im}\, g, kerg=Imf\ker g = \mathrm{Im}\, f.

Cette identité uf+vg=idu\circ f+v\circ g=\mathrm{id} est une forme abstraite du théorème de Bézout opératoriel : si deux polynômes en un endomorphisme TT, P(T)P(T) et Q(T)Q(T), vérifient une relation de Bézout A(T)P(T)+B(T)Q(T)=idA(T)P(T)+B(T)Q(T)=\mathrm{id}, alors on obtient la décomposition en somme directe E=kerP(T)kerQ(T)E=\ker P(T)\oplus\ker Q(T) (lemme des noyaux).

الحل
  1. Soit xkerfkergx\in\ker f\cap\ker g : x=(uf+vg)(x)=u(f(x))+v(g(x))=0+0=0x = (u\circ f + v\circ g)(x) = u(f(x))+v(g(x)) = 0+0 = 0. Donc kerfkerg={0}\ker f\cap\ker g=\{0\}. Pour tout xEx\in E : x=u(f(x))+v(g(x))Imuf+Imvgx = u(f(x)) + v(g(x))\in \mathrm{Im}\, u\circ f + \mathrm{Im}\, v\circ g. Comme les endomorphismes commutent, uf=fuu\circ f = f\circ u donc Im(uf)Imf\mathrm{Im}(u\circ f)\subset\mathrm{Im}\, f ; de même Im(vg)Img\mathrm{Im}(v\circ g)\subset\mathrm{Im}\, g. Donc E=Imf+ImgE=\mathrm{Im}\, f + \mathrm{Im}\, g.

  2. ker(fg)\ker(f\circ g) : décomposition x=u(f(x))+v(g(x))x = u(f(x))+v(g(x)). Si fg(x)=0f\circ g(x)=0 alors g(f(x))=f(g(x))=0g(f(x))=f(g(x))=0 (commutation). Posons x1=v(g(x))x_1 = v(g(x)) et x2=u(f(x))x_2=u(f(x)). Alors f(x1)=f(v(g(x)))=v(f(g(x)))=0f(x_1)=f(v(g(x)))=v(f(g(x)))=0 donc x1kerfx_1\in\ker f ; symétriquement g(x2)=g(u(f(x)))=u(g(f(x)))=0g(x_2)=g(u(f(x)))=u(g(f(x)))=0 donc x2kergx_2\in\ker g. Donc ker(fg)kerf+kerg\ker(f\circ g)\subset\ker f+\ker g. Inclusion inverse triviale. Somme directe car kerfkerg={0}\ker f\cap\ker g=\{0\}. Pour Im(fg)=ImfImg\mathrm{Im}(f\circ g)=\mathrm{Im}\, f\cap\mathrm{Im}\, g : évidemment Im(fg)ImfImg\mathrm{Im}(f\circ g)\subset\mathrm{Im}\, f\cap\mathrm{Im}\, g (par commutation, fg=gff\circ g=g\circ f). Réciproquement, soit yImfImgy\in\mathrm{Im}\, f\cap\mathrm{Im}\, g, y=f(a)=g(b)y=f(a)=g(b). Alors y=u(f(y))+v(g(y))=ufg(b)+vgf(a)=fg(u(b))+gf(v(a))=fg(u(b)+v(a))Im(fg)y = u(f(y))+v(g(y)) = u\circ f\circ g(b) + v\circ g\circ f(a) = f\circ g(u(b))+g\circ f(v(a)) = f\circ g(u(b)+v(a))\in\mathrm{Im}(f\circ g).

  3. Si fg=0f\circ g=0 : ker(fg)=E\ker(f\circ g)=E, donc par 2 : E=kerfkergE=\ker f\oplus\ker g. Aussi Im(fg)={0}=ImfImg\mathrm{Im}(f\circ g)=\{0\}=\mathrm{Im}\, f\cap\mathrm{Im}\, g, et par 1 : E=ImfImgE=\mathrm{Im}\, f\oplus\mathrm{Im}\, g. De plus fg=0Imgkerff\circ g=0\Rightarrow\mathrm{Im}\, g\subset\ker f ; combiné aux dimensions (dimkerf+dimImf=dimE=dimkerf+dimkergdimImf=dimkerg\dim\ker f+\dim\mathrm{Im}\, f=\dim E=\dim\ker f+\dim\ker g\Rightarrow \dim\mathrm{Im}\, f=\dim\ker g ; symétriquement dimImg=dimkerf\dim\mathrm{Im}\, g=\dim\ker f), on a kerf=Img\ker f=\mathrm{Im}\, g et kerg=Imf\ker g=\mathrm{Im}\, f.

التمرين 2

Exercice 2 — Matrice carrée alternante $A$ de taille $2n$

#matrices#rang#diagonalisation#vecteurs propres

Soient a,bRa,b\in\mathbb{R} tels que ab|a|\neq|b|. On considère la matrice carrée de taille 2n2n suivante : A=(ababbabaababbaba)A = \begin{pmatrix} a & b & a & b & \cdots \\ b & a & b & a & \cdots \\ a & b & a & b & \cdots \\ b & a & b & a & \cdots \\ \vdots & \vdots & \vdots & \vdots & \ddots \end{pmatrix} (alternée ligne par ligne à partir des entrées a,b,a,b,a,b,a,b,\dots)

  1. Calculer le rang de AA. En déduire que si n>1n>1, alors 00 est une valeur propre de AA et déterminer la dimension du sous-espace propre associé.

  2. Déterminer deux vecteurs propres associés à deux autres valeurs propres. En déduire que AA est diagonalisable.

Astuce structurelle : reconnaître que AA est de rang 2 permet de réduire l'exercice à la diagonalisation d'une matrice 2×22\times 2 sur le plan engendré par u+vu+v et uvu-v, qui sont les vecteurs propres évidents liés à la structure alternée.

الحل

Notons uR2nu\in\mathbb{R}^{2n} le vecteur (1,0,1,0,,1,0)T(1,0,1,0,\dots,1,0)^T (nn fois) et v=(0,1,0,1,,0,1)Tv=(0,1,0,1,\dots,0,1)^T (les positions impaires resp. paires). Alors on remarque que chaque ligne de AA est soit auT+bvTa\cdot u^T + b\cdot v^T (lignes impaires), soit buT+avTb\cdot u^T + a\cdot v^T (lignes paires). Donc toutes les lignes sont dans Vect(uT,vT)\mathrm{Vect}(u^T, v^T), et l'image de ATA^T est de dimension 2\le 2. Comme ab|a|\neq|b| implique (a,b)(a,b) et (b,a)(b,a) indépendants, rg(A)=2\mathrm{rg}(A)=2.

  1. Si n>1n>1, dimkerA=2n2>0\dim\ker A = 2n-2>0, donc 00 est valeur propre de multiplicité géométrique dimkerA=2n2\dim\ker A = 2n-2.

  2. Vecteurs propres non triviaux : posons x+=u+v=(1,1,1,1,)Tx_+ = u+v = (1,1,1,1,\dots)^T. Chaque ligne impaire de AA appliquée à x+x_+ donne n(a+b)n(a+b) ; chaque ligne paire aussi. Donc Ax+=n(a+b)x+Ax_+ = n(a+b) x_+, valeur propre λ1=n(a+b)\lambda_1=n(a+b). Posons x=uv=(1,1,1,1,)Tx_- = u - v = (1,-1,1,-1,\dots)^T. Ligne impaire de AA appliquée à xx_- : ab+ab+=n(ab)a-b+a-b+\dots = n(a-b) (avec nn termes aa aux positions impaires et nn termes b-b aux positions paires). Ligne paire : ba+ba+=n(ab)=n(ba)b-a+b-a+\dots = -n(a-b) = n(b-a). Donc Ax=n(ab)(1,1,1,1,)T=n(ab)xAx_- = n(a-b)\cdot(1,-1,1,-1,\dots)^T = n(a-b)\, x_-, valeur propre λ2=n(ab)\lambda_2 = n(a-b).

Comme ab|a|\ne|b|, λ1λ2\lambda_1\ne \lambda_2 et tous deux non nuls, on a trois valeurs propres 0,n(a+b),n(ab)0, n(a+b), n(a-b) avec multiplicités géométriques 2n2,1,12n-2, 1, 1, somme =2n=2n. Donc AA est diagonalisable.

التمرين 3

Exercice 3 — Applications linéaires sur $\mathbb{R}[X]_d$ : dérivation et multiplication par $X$

#applications linéaires#polynômes#matrices#noyau

Notons R[X]d\mathbb{R}[X]_d l'espace vectoriel des applications polynômiales de degré inférieur ou égal à dd. Dans ce qui suit les coordonnées seront relatives à la base canonique BdB_d de R[X]d\mathbb{R}[X]_d.

  1. Soit ff l'application de R[X]4\mathbb{R}[X]_4 dans R[X]3\mathbb{R}[X]_3 qui à une application polynômiale PP associe sa dérivée PP'. (a) Prouver que l'application ff est linéaire. (b) Donner la matrice de ff dans les bases B4B_4 et B3B_3.

  2. Soit gg l'application linéaire de R[X]3\mathbb{R}[X]_3 dans R[X]4\mathbb{R}[X]_4 définie par : g(P)=XPg(P) = XP. Donner la matrice de gg dans les bases B3B_3 et B4B_4.

  3. Donner la matrice de gfg\circ f dans la base B4B_4.

  4. Soit λR\lambda\in\mathbb{R}. Discuter suivant la valeur de λ\lambda la dimension du noyau de l'application linéaire gfλIdR[X]4g\circ f - \lambda\, \mathrm{Id}_{\mathbb{R}[X]_4}.

L'opérateur XddXX\dfrac{d}{dX} (XX fois la dérivation) est fondamental : diagonal dans la base des monômes avec valeurs propres kk (XkkXkX^k\mapsto kX^k). C'est l'opérateur d'Euler, utilisé en équations différentielles et en théorie des représentations.

الحل
  1. (a) Linéarité de la dérivation : classique. (b) f(1)=0, f(X)=1, f(X2)=2X, f(X3)=3X2, f(X4)=4X3f(1)=0,\ f(X)=1,\ f(X^2)=2X,\ f(X^3)=3X^2,\ f(X^4)=4X^3. Dans les bases B4=(1,X,X2,X3,X4)B_4=(1,X,X^2,X^3,X^4) et B3=(1,X,X2,X3)B_3=(1,X,X^2,X^3) : MB4,B3(f)=(01000002000003000004) (4 lignes, 5 colonnes).M_{B_4,B_3}(f) = \begin{pmatrix}0 & 1 & 0 & 0 & 0\\ 0 & 0 & 2 & 0 & 0\\ 0 & 0 & 0 & 3 & 0\\ 0 & 0 & 0 & 0 & 4\end{pmatrix}\ (\text{4 lignes, 5 colonnes}).

  2. g(1)=X,g(X)=X2,g(X2)=X3,g(X3)=X4g(1)=X, g(X)=X^2, g(X^2)=X^3, g(X^3)=X^4. Matrice 5×45\times 4 : MB3,B4(g)=(00001000010000100001).M_{B_3,B_4}(g) = \begin{pmatrix}0 & 0 & 0 & 0\\ 1 & 0 & 0 & 0\\ 0 & 1 & 0 & 0\\ 0 & 0 & 1 & 0\\ 0 & 0 & 0 & 1\end{pmatrix}.

  3. (gf)(P)=XP(g\circ f)(P) = X\cdot P'. Sur B4B_4 : (gf)(1)=0, X1=X, X2X=2X2, X3X2=3X3, X4X3=4X4(g\circ f)(1)=0,\ X\cdot 1 = X,\ X\cdot 2X = 2X^2,\ X\cdot 3X^2=3X^3,\ X\cdot 4X^3=4X^4. Matrice diagonale (dans B4B_4) : MB4(gf)=diag(0,1,2,3,4).M_{B_4}(g\circ f) = \mathrm{diag}(0,1,2,3,4).

  4. Les valeurs propres de gfg\circ f (donc de gfλIdg\circ f-\lambda\mathrm{Id}) sont 0,1,2,3,40,1,2,3,4, chacune de multiplicité géométrique 11 (sur le vecteur XkX^k pour k=0,1,2,3,4k=0,1,2,3,4).

• Si λ{0,1,2,3,4}\lambda\in\{0,1,2,3,4\} : dimker(gfλId)=1\dim\ker(g\circ f-\lambda\mathrm{Id}) = 1. • Sinon : gfλIdg\circ f-\lambda\mathrm{Id} est inversible et son noyau est {0}\{0\}, donc dim=0\dim = 0.