Remarks on Strichartz estimates for null forms

Fabrice Planchon

Description

Niveau: Supérieur, Doctorat, Bac+8
Remarks on Strichartz estimates for null forms Fabrice Planchon Laboratoire d'Analyse Numerique, URA CNRS 189, Universite Pierre et Marie Curie, 4 place Jussieu BP 187, 75 252 Paris Cedex Abstract We prove some improved Strichartz estimates for null forms, some of which were conjectured recently in [7]. The results follow from combining the usual Strichartz estimates with div-curl lemma techniques rather than space-time Fourier transform. Introduction Consider the bilinear form (1) Q ij (f; g) = @ i f@ j g @ i g@ j f; where f and g are functions of x 2 R n . Such forms appear in various instances connected with geometric PDEs, either elliptic ([10] and references therein) or hyperbolic ([17] and references therein). In the later context, they are called null forms (along with other bilinear forms which we will introduce later). These forms have cancellation properties. For the simple product Holder yields (2) rf;rg 2 L 2 =) @f@g 2 L 1 ; while for the null form one has (3) rf;rg 2 L 2 =) Q ij (f; g) 2 H 1 ; where H 1 denotes the Hardy space: h 2 H 1 if and only if h 2 L

• summing over dyadic


• strichartz estimates


• over


• space cancellation


• understood when considering


• rather than


• when studying


• besov space


• time fourier

Sujets

• Planchon
• Besov space

Informations

• Publié par: profil-zyak-2012
• Nombre de lectures: 29
• Langue: English

Extrait

Remarks
on
Stric
regularit
the
f
hartz
the
estimates
recast
for
L
n
g
ull
)
forms
R
F
2
abrice
f
Planc
led
hon
?
Lab
@
oratoire
f
d'Analyse
2
Num
and

i
erique,
oposition
URA
(
CNRS
j
189,
k
Univ
of
ersit
ab

_
e
r
Pierre
=
et
1
Marie
one
Curie,
2
4
f
place
where
Jussieu
2
BP
2
187,
L
75
transforms.
252
of
P
Let
aris
and
Cedex
)
Abstract
R
W
(4)
e
k
pro
2
v
view
e
an
some
follo
impro
g
v
f
ed
for
Stric
that
hartz
;
estimates
L
for
@
n
2
ull
while
forms,
ull
some
(3)
of
r
whic
2
h
ij
w
g
ere
1
conjectured
1
recen
space:
tly
1
in
if
[7
1
].
h
The
where
results
the
follo
ma
w
in
from
transforms,
com
([2
bining
;
the
2
usual
Q
Stric
;
hartz
R
estimates
j
with
f
div-curl
.
lemma
ij
tec
g
hniques
1
rather
k
than
g
space-time
:
F
the
ourier
suc
transform.
one
In
to
tro
question:
duction
f
Consider
negativ
the
,
bilinear
g
form
s
(1)
>
Q
turns
ij
has
(
f
f
r
;
2
g
2
)
)
=
f
@
g
i
L
f
;
@
for
j
n
g
form
@
has
i
r
g
;
@
g
j
L
f
=
;
Q
where
(
f
;
and
)
g
H
are
;
functions
H
of
denotes
x
Hardy
2
h
R
H
n
if
.
only
Suc
h
h
L
forms
and
app
i
ear
2
in
1
v
R
arious
denote
instances
Riesz
connected
One
with
y
geometric
(3)
PDEs,
term
either
Riesz
elliptic
Pr
([10]
1
and
])
references
f
therein)
g
or
L
h
,
yp
let
erb
ij
olic
f
([17]
g
and
=
references
i
therein).
R
In
g
the
i
later
R
con
g
text,
Then
they
kQ
are
(
called
;
n
)
ull
H
forms
.
(along
f
with
L
other
k
bilinear
k
forms
2
whic
In
h
of
w
pro
e
of
will
h
in
estimate,
tro
is
duce
naturally
later).
the
These
wing
forms
what
ha
out
v
;
e
with
cancellation
e
prop
y
erties.
namely
F
;
or
2
the
H
simple
,
pro
s
duct
0
H
It
older
out
yields
one
(2)
1Pr
oposition
2
Let
1
b
k
f
teractions
;
k
g
R
2
of
_
o
H
the
s
then
,
to
for
)
0
of
<
1
s

<
er
1
23
2
R
.
i
Then
g
(5)
k
k
k
1
(
jrj
g
2
structure
s
as
Q
the
ij
the
(
pro
f
y
;
n
g
one
)
similar
k
T
L

1
k
.
@
k
1
f
1
k

_
(1
H
ma
s

k
1
g
R
k
;
_
f
H
B
s
oth
:
and
Before
an
pro
In
ceeding
b
with
_
the
e
pro
h
of,
frequency:
w
w
e
imp
remark
of
one
ev
has
one
in
and
fact
generic
an
e
ev
,
en
]
b
adv
etter
write
result,
j
as
g
Q
i
ij

(

f
)
;
1
g
ev
)
R
2
.
_
L
B
2
2
"
s;
<
1
v
1
j
.
f
Through
B
what
+
follo
f
ws
g
the
j
reader
i
is
R
assumed
R
to
2
b
s;
e
Notice
familiar
sp
with
Q
Beso
v
v
are
spaces
as
([29]).
<
In
eac
order
four
to
individually
pro
v
v
2
e
.
(5),
left
w
term,
e
with
p
the
erform
a
a
of
parapro
distributions,
duct
in
decomp
without
osition
deca
([1]),
t
writing
Ho
(6)
there
R
instances
i
lo
f
requiremen
R
ij
j
them.
g
of
=
can
(
in
R
a
i
[5,
f
4,
;
references
R
tak
j
tage
g
w
)
R
+
f
(
k
R

j
j
g
=
;
1
R
R
i
g
f
(
)
f
+
k
(
where
R
p
i
W
f
ceed
;
k
R
k
j

g
L
)

;
f
with
k
(
k
F
2
;
s
G
~
)
Assuming
=
2
X
sum
k
high
S
(10)
k
X
2

F
j

2
k
2
G;
;
(
(
F
i
;
;
G
j
)
)
=
R
X
f
j
R
k
g
k
(
0
j
j
;
1
i

)
k
_
F
2

1
k
:
G
b
where
the

ecial
k
of
is
ij
a
the
lo
alue
calization
s
op
irrelev
erator
t
at
long
frequency
s
j
0.

fact
j
h

these
2
terms
k
elong
from
to
a
Beso
Littlew
space
o
B
o
s;
d-P
1
aley
W
decom-
are
p
with
osition,
third
and
whic
S
deals
k
in
=
at
P
same
l
for
<k
simple

duct
l
t
is
o
a
dening
lo
is
w
general
frequencies
ossible
lo
regularit
calization.
and/or
Since
y
f
the
2
w
_
distributions.
H
w
s
er,
is
are
equiv
umerous
alen
where
t
can
to
w
2
these
k
ts,
s
Q
k
is

of
k
More
f
classes
k
sym
L
ols
2
b
=
treated
"
a
k
w
2
y
l
see
2
3,
,
,
w
9
e
and
ha
therein.
v
o
e
e
(recall
an
s
of
<
structure,
0),
e
2
(9)
k
i
s
k
k
R
S

k
g
f
i
k
k
L
R
2

=
f

@
k
(
2

l
f
2
j
(and
k
similar
)
estimates
j
for
1
g
k
,
R
with

~
g
"
;
k

and
=
~


.
k
e
),
pro
from
to
whic
aluate
h

w

e
f
get
j
(7)
k
k
k
S
1
k
k
2
1
R
k
i
k
f
2


k
g
R
L
j
.
g
k
k
2
L
)
1
k
.
"
2
:
2
s
k
1
s
one

y
k
o
~
er
"
frequencies
k
get
;

and
=
then
k
(8)
1
(
k
R
R
i

f
g
;
_
R
1
j
s;
g
1
)
2and
nally
(11)
(
cs,
:
,
f
2
;
and
g
p
)
L
=
y
@
+1
i
and

u
j
stands
@
q
j
that

equation
i
k
2
v
_
(
B
the
2
e
s;
the
1

1
)
:
efore
If
p
s

=

1
n
2
u;
one
,
ma
k
y
k
sligh
can
tly
as
mo
p
dify
are
the

argumen
>
t
in
and
2
write
ma
a
estimates
p
t<