suites et series de fonctions

(Cliquez-ici pour accéder à la version originale de cette discussion avec couleurs et images)







Posted by: fonfon

Salut, je sais pas si ca a été déja posé car c'est rare que je vienne ici:

Soit\Large{(P_n)_{n\ge0}} une suite de polynômes à coefficient réels, convergeant sur R uniformément, vers une fonction f.

Prouver alors que f est un polynôme.

Ps:ce n'est pas trop dur



Posted by: abel

Y a un truc que je comprend pas :
Si on prend :
Pn = somme(k=1..n ; x^k)
Pn est bien un polyonme et pourtant il converge vers 1/(1-x) (pour |x|<1).
Enfin j'ai ptetr zappé une étape...



Posted by: fonfon

Salut, je pense que tu as peut-être zappé qq chose

moi j'utilise le fait que \Large{(P_n)_{n\ge{0}} verifie le critère de Cauchy uniforme.ET que seuls les polynomes constant sont bornés sur R....



Posted by: Bija

Oui mais ne converge pas uniformément sur R, donc ca ne correspond pas aux hypothéses du probléme posé : si on remplace R par un intervalle borné ca change tout !

Pour la solution, on a pour n>N, || Pn-PN|| <1 or les seuls polynomes bornés sont les polynomes constants donc il existe une suite de réels an tq Pn=PN+an, pour n>N. La suite an converge donc vers f(0)-PN(0)=k, mais aussi vers f(x)-PN(x), puisque Pn(x) converge vers f(x) qd n tend vers +infini, qq soit x réel.
D'ou f=PN+k est un polynôme.



Posted by: fonfon

bien Bija je vais juste redétailler un peu


La suite \Large{(P_n)_{n\ge{0}} verifie le critère de Cauchy uniforme.

Soit \Large\epsilon&gt;0; alors \Large\exists{N}\in\mathbb{N},\forall{n}\ge{N},\Large\forall{p}\in\mathbb{N},\Large\forall{x}\in\mathbb{R},|P_n(x)-P_{n+p}(x)|\le\epsilon

donc \Large\forall{p}\in\mathbb{N},\forall{x}\in\mathbb  {R},|P_N(x)-P_{N+p}(x)|\le\epsilon

Or seuls les polynômes constants sont bornés sur R, donc:
\Large\forall{p}\in\mathbb{N},\exists{C_p}\in{R} ,\Large\forall{x}\in\mathbb{R} \Large{P_N(x)+C_p=P_{N+p}(x)}.

La suite \Large{(P_{N+p}(0))_{p\in{N}} converge vers f(0), donc la suite \Large{(C_p)_{p\in{N}} converge vers \Large{f(0)}-P_N(0).

\Large\forall{x}\in\mathbb{R}, on a:

\Large{f(x)}=\lim_{p\to+\infty}P_{N+p}(x)=P_N(x)+\  lim_{p\to+\infty}C_p=P_N(x)-P_N(0)+f(0)

donc \Large\forall{x}\in\mathbb{R},f(x)=P_N(x)-P_N(0)+f(0)
donc f est bien un polynôme

A+



Posted by: atito

bien évidemment, cet exo n'est pas d'un olympiade mais juste le cours du SUP!!
Fallait penser à le poster dans supérieur je pense



Posted by: Bija

la convergence uniforme se voit en spé.











-