Global well-posedness of the generalized KP-II

equation in anisotropic Sobolev spaces

Wei Yan a , Yongsheng Li b , Yimin Zhang c
arXiv:1709.06077v1 [math.AP] 15 Sep 2017

a
School of Mathematics and Information Science and Henan Engineering Laboratory
for Big Data Statistical Analysis and Optimal Control, Henan Normal University,
Xinxiang, Henan 453007, P. R. China

b
School of Mathematics, South China University of Technology,
Guangzhou, Guangdong 510640, P. R. China
c
Department of mathematics, Wuhan University of Technology, Wuhan430070, China
Wuhan, Hubei 430070, P. R. China

Abstract. In this paper, we consider the Cauchy problem for the generalized KP-II
equation

1
ut |Dx | ux + x1 y2 u + x (u2 ) = 0, 4.
2

The goal of this paper is two-fold. Firstly, we prove that the problem is locally well-posed
in anisotropic Sobolev spaces H s1 , s2 (R2 ) with s1 > 1
4
38 , s2 0 and 4. Secondly,
we prove that the problem is globally well-posed in anisotropic Sobolev spaces H s1 , 0 (R2 )
2
with (34)
28
< s1 0. and 4. Thus, our result improves the global well-posedness
result of Hadac (Transaction of the American Mathematical Society, 360(2008), 6555-
6572.) when 4 6.

Email:yanwei19821115@sina.cn
Email:yshli@scut.edu.cn
Email: zhangyimin@whut.edu.cn
 1. Introduction

In this paper, we consider the Cauchy problem for the fifth-order KP-I equation

1
ut |Dx | ux + x1 y2 u + x (u2 ) = 0, (1.1)
2
u(x, y, 0) = u0(x, y) (1.2)

in anisotropic Sobolev space H s1 ,s2 (R2 ) defined in page 6. (1.1) occurs in the modeling
of certain long dispersive waves [1, 31, 32]. When = 2, (1.1) reduces to the KP-II
equation

1
ut + x3 u + x1 y2 u + x (u2 ) = 0. (1.3)
2

The KP-II equations arise in physical contexts as models for the propagation of dispersive
long waves with weak transverse effects [30], which are two-dimensional extensions of the
Korteweg-de-Vries equation.
Many people have investigated the Cauchy problem for KP-II equation, for instance,
see [3, 4, 9, 1519, 2129, 37, 40, 4955] and the references therein. Bourgain [4] estab-
lished the global well-posedness of the Cauchy problem for the KP-II equation in L2 (R2 )
and L2 (T2 ). Takaokao and Tzvetkov [52] and Isaza and Meja [25] established the lo-
cal well-posedness of KP-II equation in H s1,s2 (R2 ) with s1 > 31 and s2 0. Takaoka
[50] established the local well-posedness of KP-II equation in H s1,0 (R2 ) with s1 > 12
1 + 1 +
under the assumption that Dx 2 u0 L2 with the suitable chosen , where Dx 2 is
21 +
Fourier multiplier operator with multiplier || . Hadac et al. [16] established the
small data global well-posedness and scattering result of KP-II equation in the homo-
1
geneous anisotropic Sobolev space H 2 , 0 (R2 ) which can be seen in [16] and arbitrary
1
large initial data local well-posedness in both homogeneous Sobolev space H 2 , 0 (R2 )
1
and inhomogeneous anisotropic Sobolev space H 2 , 0 (R2 ).
Some author have studied the Cauchy problem for KP-I equation

1
ut x3 u + x1 y2 u + x (u2 ) = 0, (1.4)
2

for instance, see [7, 8, 14, 18, 35, 4145, 56] and the references therein. It is worth
noticing that the resonant function of KP-I equation doesnot possess the good property
as its of KP-II equation.

2
 When = 4, (1.1) reduces to the fifth-order KP-II equation

1
ut x5 u + x1 y2 u + x (u2 ) = 0. (1.5)
2

Saut and Tzvetkov [47] proved that the Cauchy problem for (1.5) is locally well-posed
in H s1 ,s2 (R2 ) with s1 > 14 , s2 0. Isaza et al. [20] proved that the Cauchy problem
for (1.5) is locally well-posed in H s1 ,s2 (R2 ) with s1 > 54 , s2 0 and globally well-posed
in H s1,0 (R2 ) with s1 > 74 . Recently, Li and Shi [38] proved that the Cauchy problem
for (1.4) is locally well-posed in H s1 ,s2 (R2 ) with s1 54 , s2 0.
Some people have studied the Cauchy problem for the fifth order KP-I equation

1
ut + x5 u + x1 y2 u + x (u2) = 0, (1.6)
2

for instance, see [6, 10, 39, 47, 48, 48] and the references therein.
In this paper, inspired by [7, 20, 39, 48], by using the Fourier restriction norm method
introduced in [2, 5, 36, 46] and developed in [33, 34], the Cauchy-Schwartz inequality and
Strichartz estimates as well as suitable splitting of domains, we prove that the Cauchy
problem for (1.1) is locally well-posed in the anisotropic Sobolev spaces H s1, s2 (R2 ) with
s1 > 32
8
and s2 0; combining the local well-posness result of this paper with
the I-method, we also prove that the problem is globally well-posed in H s1 , 0 (R2 ) with
2
s1 > (34)
28
. Thus, our result improves the result of [15].
We introduce some notations before giving the main results. Throughout this paper,
we assume that C is a positive constant which may depend upon and vary from line
to line. a b means that there exist constants Cj > 0(j = 1, 2) such that C1 |b|
|a| C2 |b|. a b means that there exist a positive constant C such that |a| > C |b|.

3
 1
0 < 1 means that 0 < < 100
. We define

hi := 1 + | |,
2
(, ) := || + ,

:= (, ), j = j (j , j )(j = 1, 2),
1
Z
F u(, , ) := 3 eixiyit u(x, y, t)dxdydt,
(2) 2 R3
1
Z
Fxy f (, ) := eixiy f (x, y)dxdy,
2 R2
1
Z
1
F u(, , ) := 3 eix+iy+it u(x, y, t)dxdydt,
(2) 2 R3
1
Z
a
Dx u(x, y, t) := 3 ||a F u(, , )eix+iy+it ddd,
(2) R
2 2

1
Z
W (t)f := eix+iy+it(,) Fxy f (, )dd.
2 R2
Let be a bump function with compact support in [2, 2] R and = 1 on (1, 1) R.
For each integer j 1, we define j () = (2j ) (21j ), 0 () = (), j (, , ) =
P
j (), thus, j () = 1. (t) is a smooth function supported in [0, 2] and equals 1 in
j0
[0, 1]. Let I R, I (x) = 1 if x I; I (x) = 0 if x doesnot belong to I.
We define
Z Z  pr ! 1r
kf kLrt Lpxy := |f |p dxdy dt .
2
R R

The anisotropic Sobolev space H s1 ,s2 is defined as follows:

n
o
s1 ,s2 2 2 s1 s2
H (R ) := u0 S (R ) : ku0 kH s1 ,s2 (R ) = khi hi Fxy u0 (, )kL2 .
2

Space Xbs1 ,s2 is defined by

 
s1 ,s2 3 s1 s2 b
Xb := u S (R ) : kukXb 1 2 = hi hi hi F u(, , ) < .
s ,s

L2 (R3 )

The space Xbs1 ,s2 ([0, T ]) denotes the restriction of Xbs1 ,s2 onto the finite time interval
[T, T ] and is equipped with the norm
n o
kukXbs1 ,s2 ([0,T ]) = inf kgkXbs1 ,s2 : g Xbs1 ,s2 , u(t) = g(t) for t [0, T ] .

For s < 0 and N N + , N 100, we define an operator IN by F IN u(, , ) =

 s
M()F u(, , ), where M() = 1 if || < N; M() = ||
N
if || N.
The main results of this paper are as follows.

4
Theorem 1.1. (Local well-posedness)

Let 4 and ||1Fxy u0 (, ) S (R2 ). Then, the Cauchy problem for (1.1) is locally
well-posed in H s1 , s2 (R2 ) with s1 > 1
4
83 , s2 0.

Remark 1. When 4 6, Hadac [15] has proved that the Cauchy problem for (1.1)
is loclly well-posed in H s1 , s2 (R2 ) with s1 > 1
4
83 , s2 0. Thus, our result extends the
result of Hadac [15].

Theorem 1.2. (Global well-posedness)

Let 4 and ||1Fxy u0 (, ) S (R2 ). Then, the Cauchy problem for (1.1) is
2
globally well-posed in H s1 , 0 (R2 ) with s1 > 3 4
12+8
.

Remark 2.When = 4, we have proved that the Cauchy problem for (1.1) is globally
well-posed in H s1 , s2 (R2 ) with s1 > 47 . Isaza and Meja [29] have proved the same
result of the Cauchy problem for (1.1) when = 4. When 4 6, Hadac [15]
has proved that the Cauchy problem for (1.1) is globally well-posed in H s1 , s2 (R2 ) with
s1 0, s2 0. Thus, our result improves the result of Hadac [15] when 4 6. Since
we can easily prove that the Cauchy problem for (1.1) is globally well-posed in H s1 , 0 (R2 )
with s1 0 with the aid of L2 conservation law of (1.1), thus, we only consider the case
2
3 4
12+8
< s1 < 0.
The rest of the paper is arranged as follows. In Section 2, we give some preliminaries.
In Section 3, we establish two crucial bilinear estimates. In Section 4, we prove the
Theorem 1.1. In Section 5, we prove the Theorem 1.2.

2. Preliminaries
In this section, motivated by [4, 45], we give Lemmas 2.1, 2.3-2.6 which play a
significant role in establishing Lemmas 3.1, 3.2. Lemma 2.2 in combination with Lemma
3.1 yields Theorem 1.1. Lemma 2.7 in combination with Lemmas 3.1, 3.2 yields Lemma
5.1.

5
Lemma 2.1. Let b > |a| 0. Then, we have that
Z b
dx 1
1 Cb , (2.1)
2

b hx + ai 2
Z
dt
ht ai
Chai , > 1, (2.2)
hti
ZR
dt 12
1 Chai , 1, (2.3)
R hti |t a| 2
Z K 1
dx K2
1 1 C 1 . (2.4)
K |x| 2 |a x| 2 |a| 2

Proof. The conclusion of (2.1) is given in (2.4) of Lemma 2.1 in [20]. (2.2)-(2.3) can
be seen in Proposition 2.2 of [48]. (2.4) can be seen in line 24 of page 6562 in [15].
This completes the proof of Lemma 2.1.

Lemma 2.2. Let T (0, 1) and s1 , s2 R and 21 < b 0 b b + 1. Then, for

h Xbs1 ,s2 , we have that

k(t)S(t)kX s1 ,s2 CkkH s1 , s2 , (2.5)

 Z t b
t
CT 1+b b khkX s1 , s2 .

S(t )h( )d (2.6)
T 0
s ,s
X 1 2
b
b

For the proof of Lemma 2.2, we refer the readers to [5, 11, 33] and Lemmas 1.7, 1.9
of [12].

Lemma 2.3. Let 2 < q , and = 12 ( 21 4 ) and b > 21 . Then,

D u
x L4t L4xy (R3 )
CkukX 0,0 . (2.7)
b

For the proof of Lemma 2.3, we refer the readers to Corollary of [15].

Lemma 2.4. Let

Z 2
Y
F P 1 (u1 , u2 )(, , ) = |1 | |2 | (1 , 1 , 1 , , , ) F uj (j , j , j )d1 d1 d1 .
4
R3 4
j=1

For b > 21 , we have

1

|Dx | 4 u2

P (u1 , u2) C |Dx | u1 . (2.8)
2

1
4 L2txy Xb0,0 Xb0,0

For the conclusion of Lemma 2.4, we refer the readers to Lemma 3.3 of [15].

6
 1
Lemma 2.5. Let b > 2
and G(1 , 1 , 1 , , , ) = f1 (1 , 1, 1 )f2 ( 1 , 1 ,
1 )f (, , ), then, we have that
2
!
1
8
Y
ku1 u2 kL2 C kDx 4
uj kX 0,0 (2.9)
b
j=1

and
 
 
Z 1 1

 |1| 4 + 8 |2 | 4 + 8 G(1 , 1 , 1 , , , ) 
d d d ddd
 
 2 1 1 1 
 R6 Q b

 hj i 
j=1
 
2
!
Y
Ckf kL2 kfj kL2 (2.10)
j=1

and
Z 
41 + 41 +
 | 1 | 8 || 8 G( 1 , ,
1 1 , , , ) 
d d d ddd
 
1 1 1
h1 ib hib
 
 R6 
2
!
Y
Ckf kL2 kfj kL2 . (2.11)
j=1

and
Z 
41 + 41 +
 || 8 | 2 | 8 G( 1 , ,
1 1 , , , ) 
d d d ddd
 
1 1 1
hib h2 ib
 
 R6 
2
!
Y
Ckf kL2 kfj kL2 . (2.12)
j=1

For the proof of Lemma 2.5, we refer the readers to Corollary 3.2 of [15].
1
Lemma 2.6. Let b > 2
and G(1 , 1 , 1 , , , ) = f1 (1 , 1, 1 )f2 ( 1 , 1 ,
1 )f (, , ), we have that
 
 
Z 
21
 |1 | |2 | G(1 , 1 , 1 , , , )
4

d1 d1 d1 ddd 
 
2

 R6 Q b

 h j i 
j=1
 
2
!
Y
Ckf kL2 kfj kL2 . (2.13)
j=1

and
Z 1

 |1 | 2 || 4 G(1 , 1 , 1 , , , ) 
d1 d1 d1 ddd 
 
b
h1 i hi b

 R6 
2
!
Y
Ckf kL2 kfj kL2 . (2.14)
j=1

7
and
Z 1

 || 2 |2| 4 G(1 , 1 , 1 , , , ) 
d d d ddd
 
b b 1 1 1
hi h2 i
 
 R6 
2
!
Y
Ckf kL2 kfj kL2 . (2.15)
j=1

and
Z 1

 || 2 |1| 4 G(1 , 1 , 1 , , , ) 
d1 d1 d1 ddd 
 
b
hi h1 i b

 R6 
2
!
Y
Ckf kL2 kfj kL2 (2.16)
j=1

and
Z 1

 |2 | 2 || 4 G(1 , 1 , 1 , , , ) 
d1 d1 d1 ddd 
 
b
h2 i hi b

 R6 
2
!
Y
Ckf kL2 kfj kL2 . (2.17)
j=1

For the proof of Lemma 2.6, we refer the readers to Proposition 3.5 of [15].

Lemma 2.7. Let () = ||, = 1 + 2 and 4 and

r (, 1) = () (1 ) (2 ). (2.18)

Then r (, 1 )12 0.

Proof.Hadac [15] and Grunrock et al. [13] have given Lemma 2.7, however, they donot
give the proof. Now we give the proof.
We consider the following eight cases:

(1) : 1 0, 2 0, 0;

(2) : 1 0, 2 0, 0;

(3) : 1 0, 2 0, 0;

(4) : 1 0, 2 0, 0;

(5) : 1 0, 2 0, 0;

(6) : 1 0, 2 0, 0.

8
We only consider cases (1), (3), (5) due to the symmetry.
When 1 = 0 or 2 = 0 or = 0, (2.12) is valid. Thus, we can assume that 1 2 6= 0.
1
Let x :=
.
When (1) is valid, we have that

r (, 1) = +1 1 x+1 (1 x)+1 .
 
(2.19)

Here 0 < x < 1. Let

F (x) = 1 x+1 (1 x)+1 . (2.20)

From (2.20), we have that

F (x) = ( + 1) [(1 x) x ] . (2.21)

When 0 < x 21 , we have that

F (x) = ( + 1) [(1 x) x ] 0. (2.22)

From (2.22), we have that

F (x) lim F (x) = 0. (2.23)

x0

1
When 2
x < 1, we have that

F (x) = ( + 1) [(1 x) x ] 0. (2.24)

From (2.24), we have that

F (x) lim F (x) = 0. (2.25)

x1

Combining (2.19) with (2.23), (2.25), we have that (2.16) is valid.

When (3) is valid, we have that

r (, 1) = +1 1 x+1 + (x 1)+1 .
 
(2.26)

Here x > 1. Let

F (x) = 1 x+1 + (x 1)+1 . (2.27)

9
From (2.27), we have that

F (x) = ( + 1) [(x 1) x ] . (2.28)

Since x > 1, we have that

F (x) = ( + 1) [(x 1) x ] < 0. (2.29)

From (2.29), we have that

F (x) lim F (x) = 0. (2.30)

x1

Combining (2.26) with (2.30), we have that (2.16) is valid.

When (5) is valid, we have that

r (, 1) = || 1 + (x)+1 (1 x)+1 .
 
(2.31)

Here x < 0. Let

F (x) = 1 + (x)+1 (1 x)+1 . (2.32)

From (2.32), we have that

F (x) = ( + 1) [(1 x) (x) ] . (2.33)

Since x < 0, we have that

F (x) = ( + 1) [(1 x) (x) ] > 0. (2.34)

From (2.34), we have that

F (x) lim F (x) = 0. (2.35)

x0

Combining (2.31) with (2.35), we have that (2.16) is valid.

This completes the proof of Lemma 2.7.

Lemma 2.8. Let = 1 + 2 , = 1 + 2 and |max | = max {||, |1|, |2|} and |min| =
min {||, |1|, |2|}. Then, we have
2
 
 ( 1 2 2 1 )
| 1 2 | = |(, ) (1, 1 ) (2 , 2 )| = r (, 1) + 
1 2 
|r (, 1)| C|max | |min|. (2.36)

10
Proof.Combining Lemma 2.6 with Lemma 3.4 of [15], we have that (2.36) is valid.
This completes the proof of Lemma 2.8.

Lemma 2.9. Let 0 < b1 < b2 < 21 . Then,

kI ()ukX 0,0 C kukX 0,0 , (2.37)

b1 b2

kI ()ukX 0,0 C kukX 0,0 . (2.38)

b2 b1

For the proof of Lemma 2.9, we refer the readers to Lemma 3.1. of [26].

3. Bilinear estimates
In this section, we give the proof of Lemmas 3.1-3.3. Lemma 3.1 is used to prove
Theorem 1.1. Lemma 3.2 in combination with I-method yields Theorems 1.2. Lemma
3.3 is used to prove Lemma 5.1.

Lemma 3.1. Let 0 < 1

100
, 4 and s1 14 83 + 4, s2 0 and uj X s11+
,s2
(j =
2

1, 2). Then, we have that

2
!
Y
kx (u1 u2 )kX s11,s2 C kuj kX s11 ,s2 . (3.1)
2 +2 2 +
j=1

Proof. To prove (3.1), by duality, it suffices to prove that

Z  2
!
  Y
 ux (u1 u2 )dxdydt CkukX s1 ,s2 kuj kX s11 ,s2 . (3.2)
1 2
R3 2 +

2 j=1

for u X s
1
1 ,s2
2
. Let
2

1
F (, , ) = his1 his2 hi 2 2 F u(, , ),
1
Fj (j , j , j ) = hj is1 hj is2 hj i 2 + F uj (j , , j )(j = 1, 2), (3.3)

and
( 2 2 2
)
X X X
6
D := (1 , 1 , 1 , , , ) R , = j , = j , = j .
j=1 j=1 j=1

To obtain (3.2), from (3.3), it suffices to prove that

2
Z ||his1 his2 F (, , )
Q
Fj (j , j , j )
j=1
2
d1 d1 d1 ddd
D 1 1
2 +
Q
hj i 2 hj is1 hj is2 hj i 2
j=1
2
!
Y
CkF kL2 kFj kL2 . (3.4)
j=1

11
Without loss of generality, by using the symmetry, we assume that |1 | |2 | and
F (, , ) 0, Fj (j , j , j ) 0(j = 1, 2) and

D := {(1 , 1, 1 , , , ) D, |1| |2 |} .

We define

1 = {(1 , 1 , 1 , , , ) D , |2| |1 | 80} ,

2 = {(1 , 1 , 1 , , , ) D , |1| 80, |1| |2 |, |2| 20},

3 = {(1 , 1 , 1 , , , ) D , |1| 80, |1| |2 |, |2| > 20},

4 = {(1 , 1 , 1 , , , ) D , |1| 80, 4|| |2 | |1|, || 20, 1 2 < 0},

5 = {(1 , 1 , 1 , , , ) D , |1| 80, 4|| |2 | |1|, || > 20, 12 < 0},

|2|
6 = {(1 , 1 , 1 , , , ) D , |1| 80, |1| |2 |, 12 < 0, || },
4
7 = {(1 , 1 , 1 , , , ) D , |1| 80, |1| |2 |, 12 > 0} .
7
Obviously, D
S
j . We define
j=1

||his1 his2
K1 (1 , 1, 1 , , , ) := 2
(3.5)
1 1
2
hj is1 hj is2 hj i 2 +
Q
hj i 2
j=1

and
Z 2
Y
Intj := K1 (1, 1 , 1 , , , )F (, , ) Fj (j , j , j )d1d1 d1 ddd
j j=1

2 2
j , we have that his2 hj is2 , thus,
P Q
with 1 j 7, j N. Since s2 0 and =
j=1 j=1
we have that

||his1
K1 (1 , 1 , 1 , , , ) 2
. (3.6)
1 1
2 +
Q s
hi 2 hj i hj i 2
1

j=1

(1). Region 1 . In this region || |1 | + |2 | 160, this case can be proved similarly to
case low + low low of pages 344-345 of Theorem 3.1 in [39].
(2). Region 2 . In this region, we have that || |1 |.

12
By using the Cauchy-Schwartz inequality with respect to 1 , 1, 1 , from (3.6), we have
that
21

||
Z Z d d d
1 1 1
Int2 C

1 2
hi 2 2

R3 R3 Q 1+2
hj i

j=1
Z 2
! 21
Y
|Fj (j , j , j )|2 d1 d1 d1 F (, , )ddd. (3.7)
3
R j=1

By using (2.2), we have that

21
Z  21
|| Z d d d
1 1 1 || d1 d1
C . (3.8)

1 2 1 1+2
hi 2 2 hi 2 2 R2 h (1 , 1 ) (2 , 2 )i

R3 Q 1+2
hj i

j=1

Let = (1 , 1) (2 , 2 ), since |1 | |2|, then we have that the absolute value

of Jacobian determinant equals
   
 (r (, 1), )   1 2 
 (1 , 1)  = 2( + 1)  1 2  ||1| |2 | |
   
 1
1  2
= 2( + 1) | + r (, 1)| 2   ||1 | |2 | |
1 2 
 1
1  2
| + r (, 1)| 
2   |1 | . (3.9)
1 2 

Inserting (3.9) into (3.8), by using (2.3), we have that

12
Z  21
|| Z d d d
1 1 1 || d1 d1
C

1 2 1 1+2
hi 2 2 hi 2 2 R2 h (1 , 1 ) (2 , 2 )i

R3 Q
hj i1+2

j=1
! 12
ddr (, 1)
Z
C
1 1
|| 2 1 hi 2 2 R2 | + r (, 1)| 2 hi1+2
! 12
dr (, 1)
Z
C
1 1 . (3.10)
|| 2 1 hi 2 2 |r (,1 )|<20|| hr (, 1) + i 2

When || < 20||, combining (3.10) with (2.1), since 4, we have that
! 12
dr (, 1)
Z
C C

1 1
2 1
1 1 C. (3.11)
|| 2 hi 2 |r (,1 )|<20|| hr (, 1) + i 2 || 4 hi 2 2

13
When || 20||, from (3.10), we have that
! 12
dr (, 1)
Z
C C|| 2
1 1 1
|| 2 1 hi 2 2 |r (,1 )|<20|| hr (, 1) + i 2 || 2 1 hi 2 2
C||
1 C. (3.12)
hi 2 2

Combining (3.8) with (3.9)-(3.12), we have that

21

|| Z d d d
1 1 1
C. (3.13)

1 2
hi 2 2

R3 Q
hj i1+2

j=1

Inserting (3.13) into (3.7), by using the Cauchy-Schwartz inequality with respect to
, , , we have that
21

||
Z Z d d d
1 1 1
Int2 C

1 2
hi 2 2

R3 R3 Q 1+2
hj i

j=1
Z 2
! 21
Y 2
|Fj (j , j , j )| d1 d1 d1 F (, , )ddd
R3 j=1

Z Z 2
! 12
Y
C |Fj (j , j , j )|2 d1 d1d1 F (, , )ddd
3 3
R R j=1

2
!
Y
CkF kL2 kFj kL2 . (3.14)
j=1

(3). Region 3 . In this region, we have that || |1 |.

Since (2.36) is valid, we have that one of the following three cases must occur:

|| := max {||, |1|, |2 |} C|1| |2 |, (3.15)

|1 | := max {||, |1|, |2 |} C|1 | |2 |, (3.16)

|2 | := max {||, |1|, |2 |} C|1 | |2 |. (3.17)

14
 1 3
When (3.15) is valid, since s1 4
8
+ 4 and 4, we have that
1
||his1 ||1 2 +2 |2 |s1 2 +2
K1 (1 , 1, 1 , , , ) 2
C 2
1 1 1
hi 2 2 hj is1 hj i 2 + hj i 2 +
Q Q
j=1 j=1
3
1 +2 58 (42) 41 + 41 +
|| 2 |2 | 8 |1 | 8 |2 | 8
C 2
C 2
. (3.18)
1 1
+ +
Q Q
hj i 2 hj i 2
j=1 j=1

Thus, combining (2.10) with (3.18), we have that

 
 2 1 
+
Z F (, , ) Q
 | j | 4 8 F ( , , )
j j j j


j=1
|Int3 | C  d d d ddd
 
2 1 1 1 
 R6 Q b

 hj i 
j=1
 
2
!
Y
kF kL2 kFj kL2 .
j=1

1 1 1 1
When (3.16) is valid, since s1 41 3
8
+ 4 and hi 2 +2 h1 i 2 hi 2 h1 i 2 +2 ,
we have that
1
||his1 ||1 2 +2 |2 |s1 2 +2
K1 (1 , 1, 1 , , , ) 2
C 1 1
1 1 h2 i 2 + hi 2 +
hi 2 2 hj is1 hj i 2 +
Q
j=1
3

58 (42) 1 1
||1 2 +2 |2 | 8 |2 | 4 + 8 || 4 + 8
C 1 1 C 1 1 . (3.19)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +
Thus, combining (2.12) with (3.19), we have that
2
!
Y
|Int3 | CkF kL2 kFj kL2 .
j=1

1 1 1 1
When (3.17) is valid, since s1 14 3
8
+4, 4, and hi 2 +2 h2 i 2 hi 2 h2 i 2 +2 ,
we have that
1
||his1 ||1 2 +2 |2 |s1 2 +2
K1 (1 , 1, 1 , , , ) 2
C 1 1
1 1 h1 i 2 + hi 2 +
hi 2 2 +
Q s
hj i hj i
1 2
j=1
3
1 +2 58 (42) 1 1
|| 2 |2 | 8 |1 | 4 + 8 || 4 + 8
C 1 1 C 1 1 . (3.20)
h1 i 2 + hi 2 + h1 i 2 + hi 2 +
Thus, combining (2.11) with (3.20), we have that
2
!
Y
|Int3 | CkF kL2 kFj kL2 .
j=1

15
(4). Region 4 . In this case, we consider (3.15)-(3.17), respectively.
1 3
When (3.15) is valid, since s1 4
8
+ 4 and 4 and || 20, we have that
||his1
K1 (1 , 1 , 1 , , , ) 2
1 1
hi 2 2 hj is1 hj i 2 +
Q
j=1
2 1
|j | 4 + 8
Q
1
+2 2s1 +2
|| 2 |2| 2 j=1
C 2
C 2
. (3.21)
1 1
hj i 2 + +
Q Q
hj i 2
j=1 j=1

Thus, combining (2.10) with (3.21), we have that

2
!
Y
|Int4 | CkF kL2 kFj kL2
j=1
1 1
When (3.16) is valid, since s1 14 3
8
+4 and 4 and || 20, hi 2 +2 h1 i 2
1 1
hi 2 h1 i 2 +2 , we have that
||his1
K1 (1 , 1 , 1 , , , ) 2
1 1
hi 2 2 hj is1 hj i 2 +
Q
j=1
1 1
+2 2s1 +2
|| 2 |2| 2 || 2 |2| 4
C 1 1 C 1 1 . (3.22)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +
Thus, combining (2.15) with (3.22), we have that
2
!
Y
|Int4 | CkF kL2 kFj kL2
j=1

When (3.17) is valid, this case can be proved similarly to case (3.16) of Region 4 with
the aid of (2.15).
(5). Region 5 .
In this region, we consider (3.15)-(3.17), respectively.
When (3.15) is valid, we have that
1
||his1 ||s1+ 2 +2 |2 |2s1 2 +2
K1 (1 , 1 , 1 , , , ) 2
C 2
; (3.23)
1 1 1
hi 2 2 + +
Q s
Q
hj i 1 hj i 2 hj i 2
j=1 j=1
1 1 3
when s1 + 2
+ 2 0, from (3.23), since s1 4
8
+ 4, we have that
2 1
|j | 4 + 8
Q
j=1
K1 (1 , 1 , 1 , , , ) C 2
; (3.24)
1
+
Q
hj i 2
j=1

16
 1 1 3
when s1 + 2
+ 2 > 0, from (3.23), since s1 4
8
+ 4, we have that
2 1
|j | 4 + 8
Q
s1 + 1 +(2+2)
|1 | 2 j=1
K1 (1 , 1 , 1 , , , ) C 2
C 2
; (3.25)
1 1
+ +
Q Q
hj i 2 hj i 2
j=1 j=1

Thus, combining (2.10) with (3.24)-(3.25), we have that

2
!
Y
|Int5 | CkF kL2 kFj kL2
j=1

1 1 1 1
When (3.16) is valid, since hi 2 +2 h1 i 2 hi 2 h1 i 2 +2 , we have that

||his1 ||his1
K1 (1 , 1 , 1 , , , ) 2
C 2
1 1 1 1
hi 2 2 hj is1 hj i 2 + hi 2 2 hj is1 hj i 2 +
Q Q
j=1 j=1
s1 + 21 +2 2s1 +2
|| |2| 2
C 1
+ 1 ; (3.26)
h2 i 2 hi 2 +
1
when s1 + 2
+ 2 21 , from (3.26), since s1 1
4
3
8
+ 4, we have that
1
|| 2 |2 | 4
K1 (1 , 1, 1 , , , ) C 1 1 ; (3.27)
hi 2 + h2 i 2 +
1
when s1 + 2
+ 2 > 12 , from (3.26), since s1 1
4
3
8
+ 4, we have that
1
||s1+1+2 || 2 |1 |2s1 2 +2
K1 (1 , 1 , 1 , , , ) C 1 1
h2 i 2 + h2 i 2 +
1 1
|| 2 |2 |s1+1 2 +(2+2) || 2 |2 | 4
C 1 1 C 1 1 ; (3.28)
hi 2 + h2 i 2 + hi 2 + h2 i 2 +

Thus, combining (2.15) with (3.27)-(3.28), we have that

2
!
Y
|Int5 | CkF kL2 kFj kL2 .
j=1

When (3.17) is valid, this case can be proved similarly to (3.16) of Region 5 with the aid
of (2.16).
(6). Region 6 .
In this region, we consider (3.15)-(3.17), respectively.

17
 1 3
When (3.15) is valid, since s1 4
8
+ 4 and 4, we have that
||his1
K1 (1 , 1 , 1 , , , ) C 2
1 1
hi 2 2 hj is1 hj i 2 +
Q
j=1
1
s1 +2(+1) 21
|2 | 2 |1 | |2| 4
C 2
C 2 . (3.29)
1 1
+ +
Q Q
hj i 2 hj i 2
j=1 j=1

Thus, combining (2.13) with (3.29), we have that

2
!
Y
|Int6 | CkF kL2 kFj kL2 .
j=1
1 1
1 3
When (3.16) is valid, since s1 4
8
+ 4 and 4 and hi 2 +2 h1 i 2
1 1
hi 2 h1 i 2 +2 , we have that
||his1
K1 (1 , 1 , 1 , , , ) C 2
1 1
hi 2 2 hj is1 hj i 2 +
Q
j=1
1 1
s1 +2(+1)
|2 | 2 |2 | 2 || 4
C 1 1 C 1 1 . (3.30)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +
Thus, combining (2.17) with (3.30), we have that
2
!
Y
|Int6 | CkF kL2 kFj kL2 .
j=1

When (3.17) is valid, this case can be proved similarly to Region 6 of (3.16) with the aid
of (2.14).
(7). Region 7 . This case can be proved similarly to Region 6 .
This ends the proof of Lemma 3.1.
1 3
Remark 3.Case (3.21) of Region 4 requires since s1 4
8
+ 4.

Lemma 3.2. Let 3

8
+ 1
2
+ 4 s < 0 and 0 < < 1
100
. Then, we have that
2
!
3
Y
kx [IN (u1 u2 ) IN u1 IN u2 ] kX 0,0 CN 4
+1+(2+2)
kIN uj kX 0,0 . (3.31)
1 +2
2 1 +
j=1 2

Proof. To prove (3.31), by duality, it suffices to prove that

Z 
 

 3 hx [I N (u 1 u 2 ) I N u 1 I N u 2 ] dxdydt

R
2
!
3
Y
+1+(2+2)
CN 4 khkX 0,0 kIN uj kX 0,0 . (3.32)
1 1 +
2 2 j=1 2

18
for h X 0,0
1
2
. Let
2

1
F (, , ) = hi 2 2 M()F h(, , ),
1
Fj (j , j , j ) = M(j )hj i 2 + F uj (j , , j )(j = 1, 2). (3.33)

To obtain (3.32), from (3.33), it suffices to prove that

2
Q
Z ||G(1, 2)F (, , ) Fj (j , j , j )
j=1
2
d1d1 d1 ddd
D 1 1
2 +
Q
hj i 2 hj i 2
j=1
2
!
3
Y
+1+(2+2)
CN 4 kF kL2 kFj kL2 , (3.34)
j=1

where

M(1 )M(2 ) M()

G(1 , 2) =
M(1 )M(2 )

and D is defined as in Lemma 3.1. Without loss of generality, we assume that F (, , )

0, Fj (j , j , j ) 0(j = 1, 2). By symmetry, we can assume that |1| |2 |.
We define
 
N
A1 = (1 , 1 , 1 , , , ) D , |2| |1| ,
2
N
A2 = {(1 , 1 , 1 , , , ) D , |1| > , |1| |2|, |2| 2},
2
N
A3 = {(1 , 1 , 1 , , , ) D , |1| > , |1| |2|, 2 < |2| N},
2
N
A4 = {(1 , 1 , 1 , , , ) D , |1| > , |1| |2|, |2| > N}.
2
4
Here D is defined as in Lemma 3.1. Obviously, D
S
Aj . We define
j=1

||G(1, 2 )
K2 (1 , 1, 1 , , , ) := 2
(3.35)
1 1
hj i 2 2 hj i 2 +
Q
j=1

and
Z 2
Y
Jk := K2 (1 , 1, 1 , , , )F (, , ) Fj (j , j , j )d1d1 d1 ddd
Aj j=1

with 1 k 4, k N.

19
Since (2.36) is valid, one of (3.15)-(3.17) must occur,
(1) Region A1 . In this case, since M(1 , 2 ) = 0, thus we have that J1 = 0.
(2) Region A2 . From page 902 of [20], we have that
|2 |
G(1 , 2) C . (3.36)
|1 |
Inserting (3.36) into (3.35) yields

||G(1, 2 ) C|2 |
K2 (1 , 1 , 1 , , , ) C 2
2
. (3.37)
1 1 1 1
2 + 2 +
Q Q
hi 2 hj i 2 hi 2 hj i 2
j=1 j=1

1
When (3.15) is valid, since 0 < < 100
, we have that
1 3 1
C|2 | 2 +2 |1| 2 +2 CN 4 +2 |2 | 2 |1| 4
K2 (1 , 1 , 1 , , , ) 2
2
. (3.38)
1 1
+ +
Q Q
hj i 2 hj i 2
j=1 j=1

By using Lemma 2.4, from (3.38), we have that

1 2
|2 | 2 |1| 4
Z
3
Y
+2
J2 CN 4
2
F (, , ) Fj (j , j , j )d1 d1 d1 ddd
D Q 1
+
hj i 2 j=1
j=1
2
!
3
Y
+2
CN 4 kF kL2 kFj kL2 . (3.39)
j=1

1 1 1 1
1
When (3.16) is valid, hi 2 +2 h1 i 2 hi 2 h1 i 2 +2 , since 0 < < 100
, we have
that
1 3 1
C|2 | 2 +2 |1 | 2 +2 CN 4
+2
|2 | 2 || 4
K2 (1 , 1 , 1 , , , ) 1 1 1 1 . (3.40)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +
By using (2.17), from (3.40), we have that
1 2
|2 | 2 || 4
Z
3
Y
+2
J2 CN 4
1 1 F (, , ) Fj (j , j , j )d1d1 d1 ddd
D h2 i 2 + hi 2 + j=1
2
!
3
Y
CN 4
+2
kF kL2 kFj kL2 . (3.41)
j=1

1 1 1 1 1
When (3.17) is valid, hi 2 +2 h2 i 2 hi 2 h2 i 2 +2 , since 0 < < 100
, we have
that
1 3

+ 12 +2 1
C|2| 2 +2 |1 | 2 +2 CN 4 |1 | 2 || 4
K2 (1 , 1 , 1 , , , ) 1 1 1 1 . (3.42)
h1 i 2 + hi 2 + h1 i 2 + hi 2 +

20
By using (2.17), from (3.42), we have that
1 2
|1 | 2 || 4
Z
3 + 12 +2
Y
J2 CN 4
1 1 F (, , ) Fj (j , j , j )d1 d1 d1 ddd
D h2 i 2 + hi 2 + j=1
2
!
3 + 12 +2
Y
CN 4 kF kL2 kFj kL2 . (3.43)
j=1

(3) Region A3 . From page 902 of [20], we have that (3.36) is valid. Combining (3.36)
with (3.35), we have that

min {||, |1|, |2|}

K2 (1 , 1 , 1 , , , ) C 2
. (3.44)
1 1
2 +
Q
hi 2 hj i 2
j=1

In this case, we consider |1 | 6|2 |, |1 | 6|2 |, respectively.

When |1 | 6|2|, we consider (3.15)-(3.17), respectively.
1
When (3.15) is valid, from (3.44), since 0 < < 100
, we have that
1 1
|2 | 2 +2 |1 | 2 +2 |2 | 2 |2 |1+2 |1 | 2 +2
K2 (1 , 1 , , ) C 2
C 2
1 1
hj i 2 + hj i 2 +
Q Q
j=1 j=1
12 +1+(2+2) 21
|2| |1 | 2 3 |2 | |1| 4
C 2
CN 4 +1+(2+2) 2 . (3.45)
1 1
hj i 2 + +
Q Q
hj i 2
j=1 j=1

By using the Plancherel identity and the Holder inequality and (2.8), from (3.45), we
have that
1
2
Z |2| 2 |1 | 4 F
Q
Fj
3 +1+(2+2) j=1
J3 CN 4
2
d1 d1 d1 ddd
D 1
+
Q
hj i 2
j=1
2
!
3
Y
+1+(2+2)
CN 4 kF kL2 kFj kL2 .
j=1

1 1 1 1 1
When (3.16) is valid, since hi 2 +2 h1 i 2 h1 i 2 +2 hi 2 and 0 < < 100
, we
have that

|2|
K2 (1 , 1 , , ) C 1
2 1 1
h1 i 2 h2 i 2 + hi 2 +
1 1
|2 | 2 +2 |1 | 2 +2 3 +1+(2+2) |2 | 2 || 4
C 1 1 CN 4
1 1 . (3.46)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +

21
Combining (2.17) with (3.46), we have that

1
2
Z |2 | 2 || 4 F
Q
Fj
3 +1+(2+2) j=1
J3 CN 4
1 1 d1d1 d1 ddd
D h2 i 2 + hi 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 .
j=1

1 1 1 1
1
When (3.17) is valid, since hi 2 +2 h2 i 2 h2 i 2 +2 hi 2 , since 0 < < 100
, we
have that

|2|
K2 (1 , 1 , , ) C 1
2 1 1
h1 i 2 h2 i 2 + hi 2 +
1
41 + 1
|2 | 2 +2 |1 | 2 +2 3 +1+(2+2) |1 |
8 || 4 + 8
C 1 1 CN 4
1 1 . (3.47)
h1 i 2 + hi 2 + h1 i 2 + hi 2 +
Combining (2.11) with (3.47), we have that
2
!
23 +6
Y
J3 CN kF kL2 kFj kL2 .
j=1

|1 |
When 6
|2 | |1 |, we consider (3.15)-(3.17), respectively.
When (3.15) is valid, from (3.44), we have that
2 1
|j | 4 + 8
Q
1
+2 +2
|| 2 |1 | 2 3 j=1
K2 (1 , 1 , , ) C 2
CN 4
+1+(2+2)
2
. (3.48)
1 1
+ +
Q Q
hj i 2 hj i 2
j=1 j=1

Combining (2.10) with (3.48), we have that

2 1
2
|j | 4 + 8 F
Q Q
Z Fj
3 +1+(2+2) j=1 j=1
J3 CN 4
2
d1d1 d1 ddd
D 1
+
Q
hj i 2
j=1
2
!
3
Y
+1+(2+2)
CN 4 kF kL2 kFj kL2 .
j=1

1 1 1 1
When (3.16) is valid, since hi 2 +2 h1 i 2 h1 i 2 +2 hi 2 , we have that

||
K2 (1 , 1 , , ) C 1
2 1 1
h1 i 2 h2 i 2 + hi 2 +
1 1
|| 2 +2 |1| 2 +2 3 +1+(2+2) || 2 |2| 4
C 1 1 CN 4
1 1 . (3.49)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +

22
Combining (2.15) with (3.49), we have that

1
2
Z || 2 |2 | 4 F
Q
Fj
3 +1+(2+2) j=1
J3 CN 4
1 1 d1d1 d1 ddd
D h2 i 2 + hi 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 .
j=1

When (3.17) is valid, this case can be proved similarly to case (3.16).
(4) Region A4 . From lines 19-20 of page 903 in [20], we have that
2  s
Y |j |
G(1 , 2 ) C . (3.50)
j=1
N

Inserting (3.50) into (3.35) yields

2
|j |s
Q
||
j=1
K2 (1 , 1 , , ) CN 2s 2
. (3.51)
1 1
2 +
Q
hi 2 hj i 2
j=1

N N
We consider || 4
, || > 4
, respectively.
N
When || 4
is valid, we consider (3.15)-(3.17), respectively.
When (3.15) is valid, from (3.51), since s 12 3
8
+ 4 and 0 < < 1
100
, we have that
1 1
+2
2s ||
2 |1|2s 2 +2 2s |1 |
2s 2 + 2 +(2+2)
K2 (1 , 1 , , ) CN 2
CN 2
1 1
hj i 2 + hj i 2 +
Q Q
j=1 j=1

3
2 1
2 1
|1 |2s 4 +1+(2+2) |j | 4 + 8 |j | 4 + 8
Q Q
j=1 3 j=1
CN 2s 2
CN 4
+1+(2+2)
2
.(3.52)
1 1
+ +
Q Q
hj i 2 hj i 2
j=1 j=1

Combining (2.10) with (3.52), we have that

2 1
2
|j | 4 + 8 F
Q Q
Z Fj
3 +1+(2+2) j=1 j=1
J4 CN 4
1 1 d1d1 d1 ddd
D h2 i 2 + hi 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 .
j=1

1 1 1 1
When (3.16) is valid, since hi 2 +2 h1 i 2 h1 i 2 +2 hi 2 and s 3
8
+ 12 + 4,

23
we have that
1
+2
2s ||
2 |1|2s 2 +2
K2 (1 , 1 , , ) CN 1 1
h2 i 2 + hi 2 +
21
2s || |1 |2s 2 +1+(2+2)
CN 1 1
h2 i 2 + hi 2 +
2s 3 +1+(2+2) 1 1
2s |1 |
4 || 2 |2 | 4 3 +1+(2+2) || 2 |2 | 4
CN 1 1 CN 4
1 1 .(3.53)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +

Combining (2.15) with (3.52), we have that

1
2
Z |2 | 2 || 4 F
Q
Fj
3 +1+(2+2) j=1
J4 CN 4
1 1 d1d1 d1 ddd
D h2 i 2 + hi 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 .
j=1

When (3.17) is valid, this case can be proved similarly to case (3.16).
|2 | |2 |
When || 6
, we consider 6
|| 6|2 |, || > 6|2|, respectively.
|2 |
When 6
|| 6|2|, we consider (3.15)-(3.17), respectively.
When (3.15) is valid, from (3.51), since s 3
8
+ 1
2
+ 4, we have that
1
+2
2s ||
2 |1|2s 2 +2
K2 (1 , 1 , , ) CN 2 1
hj i 2 +
Q
j=1

3
2 2
+ 12 +(2+2) 1 1
|1|2s |j | 4 + 8 |j | 4 + 8
Q Q
4
j=1 3
+ 12 +(2+2) j=1
CN 2s 2
CN 4
2
.(3.54)
1 1
+ +
Q Q
hj i 2 hj i 2
j=1 j=1

Combining (2.10) with (3.54), we have that

2 1
2
|j | 4 + 8 F
Q Q
Z Fj
3 +1+(2+2) j=1 j=1
J4 CN 4
1 1 d1d1 d1 ddd
D h2 i 2 + hi 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 .
j=1

1 1 1 1
When (3.16) is valid, since hi 2 +2 h1 i 2 h1 i 2 +2 hi 2 and s 3
8
+ 1
2
+

24
( + 1), we have that
1
+2
2s ||
2 |1|2s 2 +2
K2 (1 , 1 , , ) CN 1 1
h2 i 2 + hi 2 +
21
2s || |1 |2s 2 +1+(2+2)
CN 1 1
h2 i 2 + hi 2 +
2s 3 +1+(2+2) 1 1
2s |1 |
4 || 2 |2 | 4 3 +1+(2+2) || 2 |2 | 4
CN 1 1 CN 4
1 1 .(3.55)
h2 i 2 + hi 2 + h2 i 2 + hi 2 +
Combining (2.10) with (3.55), we have that
1
2
Z || 2 |2 | 4 F
Q
Fj
3 +1+(2+2) j=1
J4 CN 4
1 1 d1d1 d1 ddd
D h2 i 2 + hi 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 .
j=1

1 1 1 1
When (3.17) is valid, since hi 2 +2 h2 i 2 h2 i 2 +2 hi 2 and s 3
8
+ 12 + 4,
we have that
1
+2
2s ||
2 |1|2s 2 +2
K2 (1 , 1 , , ) CN 1 1
h1 i 2 + hi 2 +
21
2s || |1 |2s 2 +1+(2+2)
CN 1 1
h1 i 2 + hi 2 +
2s 3 +1+(2+2) 1 1
2s |1 |
4 || 2 |2 | 4 3 +1+(2+2) || 2 |2 | 4
CN 1 1 CN 4
1 1 .(3.56)
h1 i 2 + hi 2 + h1 i 2 + hi 2 +
Combining (2.10) with (3.56), we have that

1
2
Z || 2 |2 | 4 F
Q
Fj
3 +1+(2+2) j=1
J4 CN 4
1 1 d1d1 d1 ddd
D h1 i 2 + hi 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 .
j=1

When || > 6|2 |, we consider (3.15)-(3.17), respectively.

When (3.15) is valid, since s 3
8
+ 1
2
+ 4, we have that
2
|j |s
Q
||
j=1
K2 (1 , 1 , , ) CN 2s 1
2 1 1
hi 2 h1 i 2 + h2 i 2 +
21 +2s
2s |2 | |1 |1s 2 +2
CN 1 1 ; (3.57)
h1 i 2 + h2 i 2 +

25
if 2 < s, from (3.57), since s 3
8
+ 21 + 4, we have that
21 3
s+2 |2 | |1 | 4 |1 |1s 4
+2
K2 (1 , 1 , , ) CN 1 1
h1 i 2 + h2 i 2 +
1
3 +1+(2+2) |2| 2 |1 | 4
CN 4
1 1 ; (3.58)
h1 i 2 + h2 i 2 +

if 3
8
+ 21 + 4 s 2, from (3.57), since s 3
8
+ 1
2
+ 4, we have that
21 3
2s |2 | |1 | 4 |1|12s 4
+(2+2)
K2 (1 , 1 , , ) CN 1 1
h1 i 2 + h2 i 2 +
1
3 +1+(2+2) |2 | 2 |1 | 4
CN 4
1 1 , (3.59)
h1 i 2 + h2 i 2 +
combining (2.13) with (3.58)-(3.59), we have that

1
2
Z |2| 2 |1 | 4 F
Q
Fj
3 +1+(2+2) j=1
J4 CN 4
1 1 d1 d1 d1 ddd
D h1 i 2 + h2 i 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 . (3.60)
j=1

1 1 1 1
When (3.16) is valid, since hi 2 +2 h1 i 2 h1 i 2 +2 hi 2 , since s 3
8
+ 1
2
+
( + 1), we have that
2
|j |s
Q
||
j=1
K2 (1 , 1 , , ) CN 2s 1
2 1 1
h1 i 2 hi 2 + h2 i 2 +
21 +2s
2s |2 | |1 |1s 2 +2+
CN 1 1 , (3.61)
hi 2 + h2 i 2 +

if 2 < s, from (3.61), since s 3

8
+ 21 + 4, we have that
21 3
s+2 |2 | |1 | 4 |1 |1s 4
+2
K2 (1 , 1 , , ) CN 1 1
h1 i 2 + h2 i 2 +
1
3 +1+(2+2) |2 | 2 || 4
CN 4
1 1 ; (3.62)
hi 2 + h2 i 2 +

if 3
8
+ 21 + 4 s 2, from (3.61), since s 3
8
+ 1
2
+ 4, we have that
21 3
2s |2 | |1 | 4 |1|12s 4
+(2+2)
K2 (1 , 1 , , ) CN 1 1
hi 2 + h2 i 2 +
1
3 +1+(2+2) |2| 2 || 4
CN 4
1 1 , (3.63)
hi 2 + h2 i 2 +

26
combining (2.13) with (3.62)-(3.63), we have that
1
2
Z |2 | 2 || 4 F
Q
Fj
3 +1+(2+2) j=1
J4 CN 4
1 1 d1d1 d1 ddd
D hi 2 + h2 i 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 . (3.64)
j=1
1 1 1 1
When (3.17) is valid, since hi 2 +2 h2 i 2 h2 i 2 +2 hi 2 , since s 3
8
+ 12 +4,
we have that
2
|j |s
Q
||
j=1
K2 (1 , 1 , , ) CN 2s 1
2 1 1
h2 i 2 hi 2 + h1 i 2 +
21 +2s
2s |2 | |1 |1s 2 +2
CN 1 1 ; (3.65)
hi 2 + h1 i 2 +
if s 21 + 2, from (3.65), we have that
41 + 1 3
+ 32 s+2
s 12 +2 ||
8 |1| 4 + 8 |1 | 4
K2 (1 , 1 , , ) CN 1 1
hi 2 + h1 i 2 +
41 + 1
3 +1+(2+2) ||
8 |1 | 4 + 8
CN 4
1 1 ; (3.66)
hi 2 + h1 i 2 +
if 3
8
+ 21 + ( + 1) s < 12 + 2, from (3.65), we have that
41 + 1 3
+ 32 2s+2
2s ||
8 |1| 4 + 8 |1 | 4
K2 (1, 1 , , ) CN 1 1
hi 2 + h1 i 2 +
41 + 1
3 +1+(2+2) ||
8 |1 | 4 + 8
CN 4
1 1 ; (3.67)
hi 2 + h1 i 2 +
combining (2.11) with (3.66)-(3.67), we have that
1
2
Z |1 | 2 || 4 F
Q
Fj
3 +1+(2+2) j=1
J4 CN 4
1 1 d1d1 d1 ddd
D hi 2 + h1 i 2 +
2
!
3
Y
CN 4
+1+(2+2)
kF kL2 kFj kL2 . (3.68)
j=1

This completes the proof of Lemma 3.2.

Lemma 3.3. Let s 3

8
+ 1
4
+ 4, s2 0 and uj X s11+
,s2
(j = 1, 2). Then, we have
2

that
2
Y
kx I(u1 u2 )kX 0,0 C kIuj kX 0,0 . (3.69)
1
2 +2
1
2 +
j=1

27
Proof. To prove (3.69), by duality, it suffices to prove that
Z  2
!
  Y
 ux I(u1 u2 )dxdydt CkukX 0,0 kIuj kX 0,0 . (3.70)
1 2 1
R3

2 j=1 2 +

for u X 0,0
1
2
. Let
2

1
F (, , ) = hi 2 2 F u(, , ),
1
Fj (j , j , j ) = M(j )hj i 2 + F uj (j , , j )(j = 1, 2), (3.71)

and D is defined as in Lemma 3.1. To obtain (3.70), from (3.71), it suffices to prove that
2
Q
Z ||M()F (, , ) Fj (j , j , j )
j=1
2
d1 d1 d1 ddd
D 1 1
2 +
Q
hj i 2 M(j )hj i 2
j=1
2
!
Y
CkF kL2 kFj kL2 . (3.72)
j=1

From (2.4) of [26], we have that

M() his
2
C 2
. (3.73)
Q Q s
M(j ) hj i
j=1 j=1

By using (3.73), we have that the left hand side of (3.72) can be bounded by
2
Z ||hisF (, , )
Q
Fj (j , j , j )
j=1
2
d1 d1 d1 ddd. (3.74)
D 1 1
2 +
Q
hj i 2 hj is hj i 2
j=1
!
2
Q
By using (3.1), we have that (3.74) can be bounded by CkF kL2 kFj kL2 .
j=1
This completes the proof of Lemma 3.3.

4. Proof of Theorem 1.1

In this section, combining Lemmas 2.2, 3.1 with the fixed point theorem, we present
the proof of Theorem 1.1.
We define
 Z t
1 t
1 (u) := (t)W (t)u0 + W (t )x (u2 )d, (4.1)
2 0
 
B1 (0, 2Cku0kH s1 ,s2 ) := u : kukX s11 ,s2 2Cku0 kH s1 ,s2 . (4.2)
2 +

28
Combining Lemmas 2.2, 3.1 with (4.1), (4.2), we have that
 Z t
1 t 2

k1 (u)kX s1 ,s2 k(t)W (t)u0 kX s1 ,s2 + W (t )x (u )d
b 1
2 +
2 0
s1 ,s2
Xb
2



Cku0 kH s1 ,s2 + CT x (u ) s1 ,s2 X 1 +2
2

Cku0 kH s1 ,s2 + CT kuk2X s1 ,s2

b

Cku0 kH s1 ,s2 + 4C T ku0 k2H s1 ,s2 .

3
(4.3)

We choose T (0, 1) such that

1
T = 16C 2 (ku0 kH s1 ,s2 + 1)

. (4.4)

Combining (4.3) with (4.4), we have that

k1 (u)kX s1 ,s2 Cku0kH s1 ,s2 + C ku0 kH s1 ,s2 = 2C ku0 kH s1 ,s2 . (4.5)

1 +
2

Thus, 1 maps B1 (0, 2Cku0kH s1 ,s2 ) into B1 (0, 2Cku0kH s1 ,s2 ). By using Lemmas 2.2, 3.1,
(4.4)-(4.5), we have that
 Z t
1 t 2 2

k1 (u1 ) 1 (u2 )kX s1 ,s2 C
W (t )x (u1 u2 )d
1
2 +
2 0
s1 ,s2
X1
2 +
 
CT ku1 u2 kX s1 ,s2 ku1 kX s1 ,s2 + ku2 kX s1 ,s2
1 + 1 + 1
2 2 2 +
2
4C T ku0 kH s1 ,s2 ku1 u2 kX s1 ,s2
1 +
2
1
ku1 u2 kX s1 ,s2 . (4.6)
2 1
2 +

Thus, 1 is a contraction in the closed ball B1 (0, 2Cku0kH s1 ,s2 ). Consequently, u is the
fixed point of in the closed ball B1 (0, 2Cku0kH s1 ,s2 ). Then v := u|[0,T ] X s11+
,s2
([0, T ])
2

is a solution in the interval [0, T ] of the Cauchy problem for (1.1) with the initial data
u0 . For the facts that uniqueness of the solution and the solution to the Cauchy problem
for (1.1) is continuous with respect to the initial data, we refer the readers to Theorems
II, III of [24].
This ends the proof of Theorem 1.1.

5. Proof of Theorem 1.2

In this section, we give the proof of Theorem 1.2. We present the proof of Lemma
5.1 before giving the proof of Theorem 1.2.

29
Lemma 5.1. Let s1 > 14 83 and R := 1
8(C+1)3
, where C is the largest of those constants
which appear in (2.5)-(2.6), (3.69). Then, the Cauchy problem for (1.1) locally well-posed
for data satisfying IN u0 L2 (R2 ) with

kIN u0 kL2 R. (5.1)

Moreover, the solution to the Cauchy problem for (1.1) exists on a time interval [0, 1].

Proof. We define v := IN u. If u is the solution to the Cauchy problem for (1.1), then
v satisfies the following equation

1
vt + x5 v + x3 v + x1 y2 v + IN x (IN1 v)2 = 0. (5.2)
2

Then v is formally equivalent to the following integral equation

1 t
Z
v = W (t)IN u0 + W (t )IN x (IN1 v)2 . (5.3)
2 0

We define
t
1
Z
2 (v) = (t)W (t)IN u0 + (t) W (t )IN x (IN1 v)2 . (5.4)
2 0

By using Lemmas 2.2, 3.3, we have that

Z t
1 2

k2 (v)kX 0,0 k(t)W (t)IN u0 kX 0,0 + C (t)
W (t )IN x (IN v)
1 1 +
2 + 2 0 X 0,0
1
2 +

+ C IN x (IN1 v)2 X 0,0

C kIN u0 kL2
2 1 +2

+ C IN x (IN1 v)2 X 0,0

C kIN u0 kL2
1 +2
2

C kIN u0 kL2 + Ckvk2X 0,0

1 +
2

CR + Ckvk2X 0,0 . (5.5)

1
2 +

We define
 
B2 (0, 2CR) := v : kvkX 0,0 2CR . (5.6)
1 +
2

Combining (5.5)-(5.6) with the definition of R, we have that

k2 (v)kX 0,0 CR + 4C 3 R2 = 2CR. (5.7)

1
2 +

30
Thus, 2 maps B2 (0, 2CR) into B2 (0, 2CR). We define

vj = IN uj (j = 1, 2), w1 = IN1 v1 IN1 v2 , w2 = IN1 v1 + IN1 v2 . (5.8)

By using Lemmas 2.2, 3.1, 3.2, (5.5)-(5.6) and the definition of R, we have that
Z t
 1 2 1 2

k2 (v1 ) 2 (v2 )kX 0,0 C (t)
W (t )x IN (IN v1 ) (IN v2 ) d
1 +
X 0,0

2 0 1
2 +

C kx IN (w1 w2 )kX 0,0

1 +2
 2 
Ckv1 v2 kX 0,0 kv1 kX 0,0 + kv2 kX 0,0
1 + 1 1
2 2 + 2 +
1
4C 2 R2 kv1 v2 kX 0,0 kv1 v2 kX 0,0 . (5.9)
1 +
2
2 1 +
2

Thus, 2 is a contraction in the closed ball B2 (0, 2CR). Consequently, u is the fixed
point of 2 in the closed ball B2 (0, 2CR). Then v := u|[0,1] X 0,0
1
+
([0, 1]) is a solution
2

in the interval [0, 1] of the Cauchy problem for (5.3) with the initial data IN u0 . For the
uniqueness of the solution, we refer the readers to Theorem II of [24]. For the fact that
the solution to the Cauchy problem for (5.3) is continuous with respect to the initial data,
we refer the readers to Theorem III of [24]. Since the phase function (, ) is singular

at = 0, to define the derivative of W (t)u0 , the requirement ||1 Fxy u0 (, ) S (R2 )
is necessary.
This ends the proof of Lemma 5.1.
Inspired by [20], we use Lemmas 2.7, 3.2, 6.1 to prove Theorem 1.2.
For > 0, we define

u (x, y, t) = u x, 2 +1 y, +1 t , u0 (x, y) = u x, 2 +1 y .



(5.10)

Thus, u (x, y, t) X s11+

,0
([0, T ]) is the solution to
2

t u |Dx | u + x1 y2 u + u x u = 0, (5.11)

u (x, y, 0) = u0 (x, y), (5.12)

if and only if u(x, y, t) X s,0

1
+
([0, T ]) is the solution to the Cauchy problem for (1.1) in
2

[0, T ] with the initial data u0 . By using a direct computation, for (0, 1), we have
that
3
kIN u0 kL2 CN s 4
1+s
ku0 kH s,0 . (5.13)

31
For u0 6= 0 and u0 H s,0(R2 ), we choose , N such that

3 R
kIN u0 kL2 CN s 4
1+s
ku0kH s,0 := . (5.14)
4

Then there exist w3 which satisfies that kw3 kX s,0 2CR such that v := w3 |[0,1] is
1
2 +
a solution to the Cauchy problem for (5.2) with u0 . Multiplying (5.2) by 2IN u and
integrating with respect to x, y and integrating by parts with respect to x yield
Z Z
d 2
IN ux IN (u)2 dxdy = 0.
 
(IN u) dxdy + (5.15)
dt R2 R2

Combining
Z
IN ux (IN u)2 dxdy = 0
 
R2

with (5.15), we have that

Z Z
d 2
IN ux IN (u)2 (IN u)2 dxdy.


(IN u) dxdy = (5.16)
dt R2 R2

From (5.16) and Lemma 2.7, we have that

Z Z
2
(IN u(x, y, 1)) dxdy (IN u0 )2 dxdy
2 2
R R
Z 1Z
IN u x IN (u )2 (IN u )2 dxdydt


=
2
Z0 Z R
[0,1] (t)IN u [0,1] (t)x IN (u )2 (IN u )2 dxdydt




=
2
R R
C [0,1] (t)IN u 0,0 [0,1] (t)x IN (u )2 (IN u )2 0,0


X1 X 1 +
2
2
x IN (u )2 (IN u )2 0,0


C kIN u kX 0,0 X
1
2 1
2 +2
3 +1+(2+2)
CN 4 kIN u k3X 0,0 . (5.17)
1
2 +

Combining (5.14) with (5.17), we have that

R2
Z
3
(IN u(x, y, 1))2dxdy + CN 4 +1+(2+2) kIN u k3X 0,0
R2 4 1
2 +

R2 3 R2 3
+ 8C 4 N 4 +1+(2+2) R3 + 8CN 4 +1+(2+2) . (5.18)
4 4
3
Thus, if we take N sufficiently large such that such that 8C 4 N 4
+1+(2+2)
R3 34 R2 ,
then
Z  12
2
(IN u(x, y, 1)) dxdy R. (5.19)
R2

32
We consider IN u(x, y, 1) as the initial data and repeat the above argument, we obtain
that (5.11)-(5.12) possess a solution in R2 [1, 2]. In this way, we can extend the solution
to (5.11)-(5.12) to the time intyerval [0, 2]. The above argument can be repeated L steps,
where L is the maximal positive integer such that

3 3
8C 4 N 4
+1+(2+2)
R3 L R2 . (5.20)
4

More precisely, the solution to (5.11)-(5.12) can be extended to the time interval [0, L].
T
Thus, we can prove that (5.11)-(5.12) are globally well-posed in [0, +1 ] if we can choose
a number N such that

T
L . (5.21)
+1

From (5.20), we know that

3
L N 4
+1+(2+2)
. (5.22)

We know that (5.21) is valid provided that the following inequality is valid
(+1)s
3
1(2+2) T 3 1+s
CN 4 CT N 4 . (5.23)
+1

In fact, (5.23) is valid if

(+1)s
3
1 3 1+s
N 4 >N 4 (5.24)

2
which is equivalent to (34)
28
< s1 < 0.
This completes the proof of Theorem 1.2.

Acknowledgments

This work is supported by the Natural Science Foundation of China under grant
numbers 11171116 and 11401180. The first author is also supported by the Young core
Teachers program of Henan Normal University and 15A110033.

References

33
References

[1] L. A. Abramyan, Y. A. Stepanyants, The structure of two-dimensional solitons in

media with anomalously small dispersion, Sov. Phys. JETP. 61(1985), 963-966.

[2] M. Beals, Self-spreading and strength of singularities for solutions to semilinear

wave equations, Ann. of Math. 118(1983), 187-214.

[3] M. Ben-Artzi, J. C. Saut, Uniform decay estimates for a class of oscillatory integrals
and applications, Diff. Int. Eqns. 12(1999), 137-145.

[4] J. Bourgain, On the Cauchy problem for the Kadomtsev-Petviashvili equation,

Geom. Funct. Anal. 3 (1993), no. 4, 315-341.

[5] J. Bourgain, Fourier transform restriction phenomena for certain lattice subsets and
applications to nonlinear evolution equations, part II: The KdV equation, Geom.
Funct. Anal. 3(1993), 209-262.

[6] W. G. Chen, J. F. Li, C. X. Miao, On the low regularity of the fifth order Kadomtsev-
Petviashvili I equation, J. Diff. Eqns. 245(2008), 3433-3469.

[7] J. Colliander, C. E. Kenig, G. Staffilani, Low regularity solutions for the Kadomtsev-
Petviashvili I equation, Geom. Funct. Anal. 13(2003), 737-794.

[8] J. Colliander, A. D. Ionescu, C. E. Kenig, G. Staffilani, Weighted low-regularity

solutions of the KP-I initial-value problem, Discrete Contin. Dyn. Syst. 20(2008),
219-258.

[9] F. Linares, D. Pilod, J.C. Saut, The Cauchy problem for the fractional Kadomtsev-
Petviashvili equations, arXiv:1705.09744

[10] B. L. Guo, Z. H. Huo, S. M. Fang, Low regularity for the fifth order Kadomtsev-
Petviashvili-I type equation, J. Diff. Eqns. 263(2017), 5696-5726.

[11] J. Ginibre, Le probleme de Cauchy pour des EDP semi-lineaires periodiques en

variables despace (dapres Bourgain), Asterisque (1996), no. 237, Exp. No. 796, 4,
163-187, seminaire Bourbaki, Vol. 1994/95.

34
[12] A. Grunrock, New applications of the Fourier restriction norm method to wellposed-
ness problems for nonlinear Evolution Equations, Ph.D. Universitat Wuppertal,
2002, Germany, Dissertation.

[13] A. Grunrock, M. Panthee, J. D. Silva, On KP-II type equations on cylinders, Ann.

Inst. H. Poincare-AN, 26(2009), 2335-2358.

[14] Z. H. Guo, L. Z. Peng, B. X. Wang, On the local regularity of the KP-I equation in
anisotropic Sobolev space, J. Math. Pures Appl. 94(2010), 414-432.

[15] M. Hadac, Well-posedness for the Kadomtsev-Petviashvili II equation and general-

izations, Trans. Ameri. Math. Soc. 360(2008), 6555-6572.

[16] M. Hadac, S. Herr, H. Koch, Well-posedness and scattering for the KP-II equation
in a critical space, Ann. Inst. H. Poincare-AN, (2009), 917-941.

[17] N. Hayashi, P. l. Naumkin, Large time asymptotics for the Kadomtsev-Petviashvili

equation, Comm. Math. Phys. 332(2014), 505-533.

[18] A. D. Ionescu, C. E. Kenig, D. Tataru, Global well-posedness of the KP-I initial-

value problem in the energy space, Invent. Math. 173(2008), 265-304.

[19] P. Isaza, F. Linares, G. Ponce, On the propagation of regularity of solutions of the

Kadomtsev-Petviashvili equation, SIAM J. Math. Anal. 48(2016), 1006-1024.

[20] P. Isaza, J. Lopez, J. Meja, Cauchy problem for the fifth order Kadomtsev-
Petviashvili (KPII) equation, Commun. Pure Appl. Anal. 5(2006), 887-905.

[21] P. Isaza, J. Meja, V. Stallbohm, Local solution for the Kadomtsev-Petviashvili

equation with periodic conditions, Manuscripta Math. 75(1992), 383-393.

[22] P. Isaza, J. Meja, V. Stallbohm, A regularity theorem for the Kadomtsev-

Petviashvili equation with periodic boundary conditions, Nonlinear Anal. 23(1994),
683-687.

[23] P. Isaza, J. Meja, V. Stallbohm, Local solution for the Kadomtsev-Petviashvili

equation in R2 , J. Math. Anal. Appl. 196(1995), 566-587.

35
[24] P. Isaza, J. Meja V. Stallbohm, The Cauchy problem for the Kadomtsev-
Petviashvili equations (KPII) in Sobolev spaces H s , s > 0, Diff. Int. Eqns. 14(2001),
529-557.

[25] P. Isaza, J. Meja, Local and global Cauchy problems for the Kadomtsev-Petviashvili
(KP-II) equation in Sobolev spaces of negative indices, Comm. Partial Diff. Eqns.
26(2001), 1027-1054.

[26] P. Isaza, J. Meja, Global solution for the Kadomtsev-Petviashvili equation (KP-II)
in anisotropic Sobolev spaces of negative indices, Electr. J. Diff. Eqns. 2003, 68, 12
pp.

[27] P. Isaza, J. Meja, N. Tzvetkov, A smoothing effect and polynomial growth of the
Sobolev norms for the KP-II equation, J. Diff. Eqns. 220(2006), 1-17.

[28] P. Isaza, J. Lopez, J. Meja, The Cauchy problem for the Kadomtsev-Petviashvili
(KPII) equation in three space dimensions, Comm. Partial Diff. Eqns. 32(2007),
611-641.

[29] P. Isaza, J. Meja, On the support of solutions to the Kadomtsev-Petviashvili (KP-

II) equation, Commun. Pure Appl. Anal. 10(2011), 1239-1255.

[30] B. B. Kadomtsev, V. I. Petviashvili, On the stability of solitary waves in weakly

dispersive media, Soviet. Phys. Dokl. 15(1970), 539-541.

[31] V. I. Karpman, V. Yu. Belashov, Dynamics of two dimensional solitons in weakly

dispersive media, Phys. Lett. A. 154(1991), 131-139.

[32] V. I. Karpman, V. Yu. Belashov, Evolution of three-dimensional nonlinear pulses

weakly dispersive media, Phys. Lett. A. 154(1991), 1401-144.

[33] C. E. Kenig, G. Ponce, L. Vega, The Cauchy problem for the Korteweg-de Vries
equation in Sobolev spaces of negative indices, Duke Math. J. 71(1993), 1-21.

[34] C. E. Kenig, G. Ponce, L. Vega, A bilinear estimate with applications to the KdV
equation, J. Amer. Math. Soc. 9(1996), 573-603.

36
[35] C. E. Kenig, On the local and global well-posedness theory for the KP-I equation,
Ann I. H. Poincar-AN, 21(2004), 827-838.

[36] S. Klainerman, M. Machedon, Smoothing estimates for null forms and applications,
Int. Math. Res. Not. (9):383ff. approx. 7 pp. (electronic), 1994.

[37] H. Koch, J. F. Li, Global well-posedness and scattering for small data for the
three-dimensional Kadomtsev-Petviashvili II equation, Commun. Partial Diff. Eqns.
42(2017), 950-976.

[38] J. F. Li, S. G. Shi, A local well-posed result for the fifth order KP-II initial value
problem, J. Math. Anal. Appl. 402(2013), 679-692.

[39] J. F. Li, J. Xiao, Well-posedness of the fifth order Kadomtsev-Petviashvili I equa-

tion in anisotropic Sobolev spaces with nonnegative indices, J. Math. Pures Appl.
90(2008), 338-352.

[40] Y. Liu, Strong instability of solitary-wave solutions to a Kadomtsev-Petviashvili

equation in three dimensions, J. Diff. Eqns. 180(2002), 153-170.

[41] L. Molinet, J. C. Saut, N. Tzvetkov, Well-posedness and ill-posedness results for

the Kadomtsev-Petviashvili-I equation, Duke Math. J. 115(2002), 353-384.

[42] L. Molinet, J. C. Saut, N. Tzvetkov, Global well-posedness for the KP-I equation,
Math. Ann. 324(2002), 255-275.

[43] L. Molinet, J. C. Saut, N. Tzvetkov, Correction: Global well-posedness for the

KP-I equation [Math. Ann.324(2002), 255-275;MR1933858] Math. Ann. 328(2004),
707-710.

[44] L. Molinet, J. C. Saut, N. Tzvetkov, Global well-posedness for the KP-I equation on
the background of a non localized solution, Comm. Math. Phys. 272(2007), 775-810.

[45] L. C. Molinet, J. C. Saut, N. Tzvetkov, Global well-posedness for the KP-II equa-
tion on the background of a non-localized solution, Ann. Inst. H. Poincare-AN, 28
(2011), 653-676.

37
[46] J. Rauch, M. Reed, Nonlinear microlocal analysis of semilinear hyperbolic systems
in one space dimension, Duke Math. J. 49(1982), 397-475.

[47] J. C. Saut, N. Tzvetkov, The Cauchy problem for higher-order KP equations, J.

Diff. Eqns. 153(1999), 196-222.

[48] J. C. Saut, N. Tzvetkov, The Cauchy problem for the fifth order KP equations, J.
Math. Pures Appl. 79(2000), 307-338.

[49] J. C. Saut, N. Tzvetkov, On periodic KP-I type equations, Comm. Math. Phys.
221(2001), 451-476.

[50] H. Takaoka, Global well-posedness for the Kadomtsev-Petviashvili II equation, Dis-

crete Contin. Dyn. Syst. 6(2000), 483-499.

[51] H. Takaoka, Well-posedness for the Kadomtsev-Petviashvili II equation, Adv. Diff.

Eqns. 5(2000), 1421-1443.

[52] H. Takaoka, N. Tzvetkov, On the local regularity of the Kadomtsev-Petviashvili-II

equation, Int. Math. Res. Not. 2001(2001), 77-114.

[53] N. Tzvetkov, On the Cauchy problem for Kadomtsev-Petviashvili equation, Comm.

Partial Diff. Eqns. 24(1999), 1367-1397.

[54] N. Tzvetkov, Global low-regularity solutions for Kadomtsev-Petviashvili equation,

Diff. Int. Eqns. 13(2000), 1289-1320.

[55] N. Tzvetkov, Long time bounds for the periodic KP-II equation, Int. Math. Res.
Not. 2004(2004), 2485-2496.

[56] Y. Zhang, Local well-posedness of KP-I initial value problem on torus in the Besov
space, Comm. Partial Diff. Eqns. 41(2016), 256-281.

38