23.07.2013 Views

Asymptotic behaviour of the Kazdan-Warner solution in the annulus ∗

Asymptotic behaviour of the Kazdan-Warner solution in the annulus ∗

Asymptotic behaviour of the Kazdan-Warner solution in the annulus ∗

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

(see Section 2 for some remarks on Ga,b(r, s)). Note that us<strong>in</strong>g <strong>the</strong> Green’s<br />

function we can write Theorem 1.1 <strong>in</strong> <strong>the</strong> follow<strong>in</strong>g way,<br />

Corollary 1.2. Let up <strong>the</strong> unique radial <strong>solution</strong> <strong>of</strong> (1.1). Then, as p → ∞,<br />

where τ is given by<br />

up(|x|) → τGa,b(|x|, r0) <strong>in</strong> C 0 (Ω). (1.6)<br />

⎧<br />

⎪⎨<br />

τ =<br />

⎪⎩<br />

4(N−2)<br />

N −1<br />

r0 (a2−N −b2−N )<br />

4<br />

r0(log b−log a)<br />

if N > 2<br />

if N = 2<br />

(1.7)<br />

From Theorem 1.1 we deduce <strong>the</strong> follow<strong>in</strong>g sharp Sobolev <strong>in</strong>equality for<br />

radial functions <strong>in</strong> <strong>the</strong> <strong>annulus</strong>,<br />

Theorem 1.3. Let Ω be <strong>the</strong> <strong>annulus</strong> Ω = {x ∈ R N : 0 < a < |x| < b}. Then,<br />

for any radial function u ∈ H 1 0(Ω) <strong>the</strong> follow<strong>in</strong>g <strong>in</strong>equality holds,<br />

where<br />

<br />

|∇u|<br />

Ω<br />

2 ≥ Cp<br />

Cp →<br />

<br />

u<br />

Ω<br />

p<br />

2<br />

p<br />

<br />

ωN 4(N−2)<br />

a2−N −b2−N if N ≥ 3<br />

8π<br />

log b−log a if N = 2<br />

, for any p > 1 (1.8)<br />

as p → ∞. Here ωN denotes <strong>the</strong> area <strong>of</strong> <strong>the</strong> unit sphere <strong>in</strong> R N .<br />

(1.9)<br />

Observe that Theorem 1.1 implies that ||up||∞ → 1. Next results gives a<br />

more precise estimate.<br />

Theorem 1.4. The follow<strong>in</strong>g estimate holds<br />

1<br />

where γ = lim 2 r→r0<br />

ω′ (r) 2 =<br />

||up||∞ = 1 +<br />

log p<br />

p<br />

⎧ <br />

⎪⎨<br />

log (N − 2) 22 2<br />

⎪⎩<br />

<br />

log<br />

2<br />

ab(log b−log a) 2<br />

<br />

γ 1<br />

+ + o<br />

p p<br />

2−N (a2−N +b 2−N N −1<br />

2<br />

) N −2<br />

(a2−N −b2−N ) 2<br />

<br />

<br />

(1.10)<br />

if N ≥ 3<br />

if N = 2<br />

We po<strong>in</strong>t out that Theorems 1.3 and 1.4 are proved us<strong>in</strong>g <strong>the</strong> ”global” convergence<br />

result <strong>in</strong> Theorem 1.1. Moreover <strong>the</strong> pro<strong>of</strong> <strong>of</strong> <strong>the</strong>se results just use<br />

elementary arguments.<br />

The limit function ω(r) <strong>in</strong> Theorem 1.1 is not differentiable at r = r0.<br />

Actually, it is <strong>in</strong>terest<strong>in</strong>g to study more carefully <strong>the</strong> <strong>behaviour</strong> <strong>of</strong> <strong>the</strong> <strong>solution</strong><br />

up(r) near <strong>the</strong> maximum r0. This leads to analyze <strong>the</strong> ”local” convergence <strong>of</strong><br />

<strong>the</strong> <strong>solution</strong> up(r) near its maximum. In order to do this we use a blow-up<br />

3

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!