• Aucun résultat trouvé

(b) Surface element of Euclidean spheres Show that the degreen−1 differential form onRn defined by dσ(x

N/A
N/A
Protected

Academic year: 2022

Partager "(b) Surface element of Euclidean spheres Show that the degreen−1 differential form onRn defined by dσ(x"

Copied!
2
0
0

Texte intégral

(1)

M2R Université Grenoble Alpes 2019-2020 Introduction to analytic geometry (course by Jean-Pierre Demailly)

Sheet number 1, 03/10/2019

1. Standard area and volume calculations in Rn and Cn. Here, Rn and Cn are equipped with their standard Euclidean/Hermitian structures, the Hermitian norm being denoted z7→ |z|.

(a) The Lebesgue measure inCn'R2n can be written as

dλ(z) = i

2dz1∧dz1∧. . .∧ i

2dzn∧dzn= in2

2n dz1∧. . .∧dzn∧dz1∧. . .∧dzn,

wherezj =xj+iyn andCn 'R2n is oriented (conventionally) by coordinates(x1, y1, . . . , xn, yn).

(b) Surface element of Euclidean spheres

Show that the degreen−1 differential form onRn defined by

dσ(x) = 1 R

n

X

j=1

(−1)j−1xjdx1∧. . .∧ddxj∧. . .∧dxn (where b∗ means a term∗ omitted),

induces on the sphereS(0, R)⊂Rn the Euclidean area measure.

In the case of Cn, show that a similar role can be played by

dσ(z) = in2 2n

2 R

n

X

j=1

(−1)j−1zjdz1∧. . .∧dzcj ∧. . .∧dzn∧dz1∧. . .∧dzn.

Hint. Computedr(x)∧dσ(x) withr(x) = (P

|xj|2)1/2 (resp. r(z) = (P

zjzj)1/2); compare with dλ.

(c) Ifx=ruis the usual polar decomposition inRnr{0} (withr =|x| ∈R+ and u= |x|x ∈Sn−1, show that dλ(x) =rn−1dr∧dσ(u) where dσ(u) is given as in (b). (Of course, one has a similar formula inCn withz=ru and dλ(z) =r2n−1dr∧dσ(u)).

Hint. Notice thatdu1∧. . .∧dun= 0 sincedimSn−1< n.

(d) LetS2n−1⊂Cn be the unit sphere, and consider the projection

S2n−1→Cn−1, z= (z1, . . . , zn)7→z0= (z1, . . . , zn−1).

Writez0=ρu,ρ∈R+,u∈S2n−3⊂Cn−1, and zn =rnen in polar coordinates on Cn−1 (resp. C).

Ifλ0 is the Lebesgue measure ofCn−1and σ0 the area measure ofS2n−3, show that

dσ(z) =dλ0(z0)∧dθn2n−3dρ∧dσ0(u)∧dθn

in the complement ofS2n−1∩ {zn= 0} 'S2n−3(which is dσ-negligible) inS2n−1. Derive from there the Fubini type formula

Z

S2n−1

f(z)dσ(z) = Z

ρ∈[0,1]

Z

θn∈[0,2π]

Z

u∈S2n−3

f ρu,(1−ρ2)1/2en

ρ2n−3dρ dθn0(u).

Hint. Usen= Im(dlogzn) = 2i1(dzzn

ndzzn

n ) to compute dr∧dλ(z0)∧dθn.

(e) Show by induction on n that the area of S2n−1 is σ2n−1 = (n−1)!n . The area of S(0, R) ⊂ Cn is

n

(n−1)!R2n−1and the volume ofB(0, R)⊂Cn is πn!nR2n. 2. LetΩ⊂Cn be an open set such that

∀z∈Ω, ∀λ∈C, |λ| ≤1 ⇒ λz ∈Ω.

Show that Ω is a union of polydisks of center 0 (with respect to coordinates z0 = u(z) associated with arbitrary unitary matricesu∈U(n)) and infer that the space of polynomialsC[z1, . . . , zn]is dense inO(Ω)

1

(2)

for the topology of uniform convergence on compact subsets. IfΩ is bounded, show thatC[z1, . . . , zn]is dense inO(Ω)∩C0(Ω)for the topology of uniform convergence on Ω.

Hint: consider the Taylor expansion of a function f ∈ O(Ω) at the origin, writing it as a series of homogeneous polynomials. To deal with the case of O(Ω)∩C0(Ω), first apply a dilation tof.

3. The goal of this exercise is to prove the Cauchy formula for the unit ball inCn. Let B ⊂Cn be the unit Hermitian ball,S=∂B andf ∈O(B)∩C0(B). Our goal is to check the following Cauchy formula:

f(w) = 1 σ2n−1

Z

S

f(z)

(1− hw, zi)n dσ(z).

(a) By means of a unitary transformation and exercise 2, reduce the question to the case when w = (0, . . . ,0, wn)and f(z) is a monomialzα=z1α1. . . znαn.

(b) Show that the integralR

Szαzkndσ(z) vanishes unlessα= (0, . . . ,0, k). Compute the value of the non zero integral by means of suitable integration by parts.

Hint. Use formula 1 (d), invariance ofσ by rotation, and/or Fourier series arguments.

(c) Prove the formula by means of a suitable power series expansion of(1− hw, zi)−n. 4. Montel spaces

Let E be a topological vector space E over K =R or C. A subset S ⊂ E is said to be bounded if for every neighborhoodU of zero inE, there existsλ∈K such thatA⊂λU (or equivalently, λ−1A⊂U).

(a) Show that ifA is compact, then A is bounded.

Hint. Use the continuity of the operations to prove the fact that for any given open neighborhoodU of0, one can find a neighborhoodV of0 andδ >0such thatV +µV ⊂U for allµ∈K,|µ|< δ, and coverA by translatesxj+V.

(b) If (E,k k) is a normed vector space, show that A is bounded if and only if supx∈Akxk < +∞.

More generally, ifE is locally convex and the topology ofE is defined by a collection of semi-norms (pα)α∈I, thenA is bounded if and only if for everyα∈I one hassupx∈Apα(x)<+∞.

(c) A Fréchet spaceEwill be said to be aMontel spaceif the compact subsets ofE are exactly the closed bounded subsets of E. Show that for every open set Ω ⊂ Cn, the space O(Ω) is a Montel space;

likewise, forΩ⊂Rn,C(Ω)is a Montel space.

(d) Show that an infinite dimensional Banach space is never a Montel space. Derive from there that the topology of O(Ω)orC(Ω)cannot be defined by a single norm.

2

Références

Documents relatifs

Beyond general statements, the ideal of transparency is materialized through private reporting standards such as the Global Reporting Initiative (GRI), the

random vec- tors which use our identities to provide rates of convergence in information theoretic central limit theorems (both in Fisher information distance and in relative

According to the values found at the preceding question, discuss and decide (if possible) on the nature of the critical points (extremum, saddle points ?).. Write the expression of

Description of the absorption spectrum of iodine recorded by means of Fourier Transform Spectroscopy :.. the

Table VII is useful for people who need to identify a few transitions as frequently occurs in laser spectroscopy, and also enables one to check the

We prove a local version of a (global) result of Merle and Zaag about ODE behavior of solutions near blowup points for subcritical nonlinear heat equations.. This solves a problem

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des

For agreeableness - (information considered unpleasant) and still at the positive pole of the dimensions, the following hierarchies are noticed systematically (i.e. both for the