09.01.2014 Views

Function Spaces as Dirichlet Spaces (About a Paper by Maz'ya and ...

Function Spaces as Dirichlet Spaces (About a Paper by Maz'ya and ...

Function Spaces as Dirichlet Spaces (About a Paper by Maz'ya and ...

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.

16 N. Jacob <strong>and</strong> R. L. Schilling<br />

Again we use that s ↦→ g(s)/s is decre<strong>as</strong>ing to find<br />

∫ ∞<br />

1<br />

x −n/2−1 h(x) dx<br />

c ′ n<br />

{ ∫ ∞<br />

1<br />

1<br />

[ ∫ 1<br />

2<br />

0<br />

g( x 2 )<br />

x<br />

2<br />

] ∫ dx ∞<br />

[ ∫ 1<br />

t n/2−1 dt<br />

x +<br />

{( ∫ ∞<br />

g( x<br />

˜c ) )( ∫ 1 )<br />

2<br />

2<br />

n ( x 2<br />

dx t<br />

2) n/2−1 dt +<br />

0<br />

1<br />

1<br />

2<br />

∫ ∞<br />

1<br />

g( x 2 )<br />

x<br />

2<br />

( ∫ 1<br />

2<br />

0<br />

] } dx<br />

dt<br />

x<br />

g(xs)<br />

xs<br />

ds ) dx<br />

x<br />

}<br />

.<br />

By <strong>as</strong>sumption, the first term is finite, <strong>and</strong> for the second term we find<br />

∫ ∞<br />

( ∫ 1<br />

( ∫<br />

2<br />

∞<br />

1<br />

<strong>as</strong> µ( 1 ) µ(1) < ∞.<br />

2<br />

0<br />

) ∫ 1<br />

g(xs) dx<br />

xs ds x = 2<br />

0<br />

s<br />

g(y)<br />

y<br />

dy<br />

y<br />

)<br />

ds = µ ( 1<br />

2)<br />

< ∞<br />

Corollary 4.2. The Bernstein function f <strong>as</strong>sociated with s ↦→ g(s −1 )s −n/2 is<br />

given <strong>by</strong><br />

f(r) = 1 π n/2 ∫ ∞<br />

(<br />

4 Γ ( 1 − e<br />

)<br />

−rλ ∫ ∞<br />

[ ∫ 1 (<br />

exp − 1 − t ) ] ) ρ(ds)<br />

t n/2−1 dt dλ .<br />

n<br />

2 0 λ 2 0+ 0 4sλ<br />

s<br />

Proof. Applying Lemma 2.1 to the me<strong>as</strong>ure η(dx) = h(x) dx, x > 0, we find<br />

f(r) =<br />

∫ ∞<br />

0+<br />

(1 − e −rx )(4πx) n/2 Φ(η)(dx)<br />

where Φ(x) = 1/4x. Hence<br />

∫ ∞<br />

( ∫ ∞<br />

[ ∫ 1<br />

] )<br />

f(r) = πn/2<br />

ρ(ds)<br />

Γ ( )<br />

n<br />

(1 − e −r/4x ) x −n/2 x n/2 e −x/s t n/2−1 e −tx/s dt dx<br />

2 0<br />

0+<br />

0<br />

s<br />

∫ ∞<br />

( ∫ ∞<br />

[ ∫ 1<br />

] )<br />

= πn/2<br />

ρ(ds)<br />

Γ ( )<br />

n<br />

(1 − e −r/4x )e −x(1−t)/s t n/2−1 dt dx<br />

2 0 0+ 0<br />

s<br />

∫ ∞<br />

( ∫ ∞<br />

[ ∫ 1<br />

= πn/2<br />

(<br />

4 Γ ( )<br />

n<br />

(1 − e −rλ ) exp − 1 − t ) ] ) ρ(ds) dλ<br />

t n/2−1 dt<br />

2 0 0+ 0<br />

4sλ<br />

s λ 2 ,<br />

where we used the substitution λ = 1/4x.<br />

It is <strong>by</strong> no means obvious to find out the c<strong>as</strong>es where f <strong>and</strong> µ are equal.<br />

However, we can always prove that f ∼ µ holds.<br />

Proposition 4.3. Assume that f is <strong>as</strong> in the above corollary, µ <strong>as</strong> in Lemma 3.3<br />

<strong>and</strong> that the <strong>as</strong>sumptions of Lemma 4.1 are satisfied. Then there are two constants<br />

κ 0 , κ 1 > 0 such that<br />

κ 0 f(r) µ(r) κ 1 f(r), r > 0.

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

Saved successfully!

Ooh no, something went wrong!