Advances in Pure Mathematics, 2011, 1, 351-358
doi:10.4236/apm.2011.16063 Published Online November 2011 (http://www.SciRP.org/journal/apm)
Copyright © 2011 SciRes. APM
Existence and Nonexistence of Entire Positive Solutions for
a Class of Singular p-Laplacian Elliptic System*
Daojin Lei1, Zuodo ng Yang1,2
1School of Mathematical Sciences, Nanjing Normal University, Nanjing, China
2College of Zhongbei, Nanjing Normal University, Nanjing, China
E-mail: zdyang_jin@263.net
Received August 7, 2011; revised October 2, 2011; accepted October 10, 2011
Abstract
In this paper, we show the existence and nonexistence of entire positive solutions for a class of singular el-
liptic system




 
2
2
=,
=,, 3
ap p
aq q
divxuubxfu v
divxvvdxgu vxRN



 
N.
We have that entire large positive solutions fail to exist if f and g are sublinear and b and d have fast decay at
infinity. However, if f and g satisfy some growth conditions at infinity, and b, d are of slow decay or fast de-
cay at infinity, then the system has infinitely many entire solutions, which are large or bounded.
Keywords: Singular p-Laplacian Elliptic System, Entire Positive Solution, Large Solution, Bounded
Solution, Entire Large Positive Radial Solution
1. Introduction
In this paper, we mainly consider the existence and
nonexistence of positive solutions for the following
singular p-laplacian elliptic system:






2
2
=,,
=,, ,
ap p ,
N
aq q
N
divxuubxfuvx
divxvvdxguvx


 
 
R
R (1)
where , and are continuous, posi-
tive and nondecreasing functions in
3,> 0Nab d
N
R,
are positive, nondecreasing
and continuous functions.
, :)fg  [0,)[0,) [0,
When , the following semi-linear
elliptic system:
=0,==2apq


,,=0, in ,
,,=0, in ,
ufxuv
vgxuv
 
 
has been studied extensively over the years, for
example see [1-4]. If

=,=
f
bxv gdxu
, the
above system becomes

=, ,
=, ,
N
N
ubxvx
vdxux


R
R
for which existence results for boundary blow-up posi-
tive solution can be found in a recent paper by Lair and
Wood [5]. The authors established that all positive entire
radial solutions of system above are boundary blow-up
provided that
 
00
d=, d=.tbt ttdt t


On the other hand, if
 
00
d<, d<,tbt ttdt t


then all positive entire radial solutions of this system are
bounded.
F. Cìrstea and V. Rădulescu [2] extended the above
results to a larger class of systems



=, ,
=,
N
,
N
ubxgvx
vdxfux


R
R
*Project Supported by the National Natural Science Foundation o
f
China (No.11171092); the Natural Science Foundation of the Jiangsu
Higher Education Institutions of China (No.08KJB110005).
D. J. LEI ET AL.
352
Z. D.Yang [6] extended the above results to a class of
quasi-linear elliptic system:




2
2
=, ,
(||)=(||), ,
pN
q
divuubxg vx
divvvdxf ux
 
 
R
RN
where , , and
3N,>1pq
,
N
bd CR are
positive functions,
0,)
1[C,fg
f are positive and
non-decreasing. When and
g
satisfy
(H1)
 

0=0=0,= >0;
lim infu
fu
fg gu

(H2) The Keller-Osserman condition,
  
10
d<, =d,
t
p
tGtgs s
Gt

then there exists an entire positive radial solution, and in
addition, the function satisfy
,bd
(H3)

1
1
11
00 dd=
tp
NN
tsbss t

 ;
(H4)

1
1
11
00 dd=
tq
NN
tsdss t

 ,
then all entire positive radial solutions are large.
On the other hand, if and satisfy
b d
(H5)

1
1
11
00 dd<
tp
NN
tsbss t

 ;
(H6)

1
1
11
00 dd<
tq
NN
tsdss t

 ,
then all entire positive radial solutions are bounded.
While in [7], the author got the relevant results of the
same system only on the following conditions
(H7) are continuous;
, , ,:[0,)[0,)bdg f 
(H8)
g
and
f
are non-decreasing functions on
[0, ).
However, when , there are few results
about the existence and nonexistence of singular p-La-
placian elliptic system (1). And as to the single equation,
we can refer to [8]. The present results are complements
and extensions of some results in [7,9], which to be more
precise, if , you can get the relevant
existence and nonexistence results for a class of semi-
linear elliptic system with gradient term in [9]; Mean-
while, if , you can get the relevant
existence results for a class of quasi-linear elliptic system
in [7].
0, 2ap
= =2q
, >1
q
=0,ap
=0,ap
For convenience, we need the following definition:
Definition. A solution of system (,)uv




2
2
=,,, ,
=,,, ,
ap p
aq q
divxuufx uvx
divxvvgx uvx



 
(2)
is called an entire large solution(or explosive solution, or
blow-up solution), if it is classical solution of (2) on
and
N
R
ux and
vx as x.
Now we give our main theorem:
Theorem 1. Suppose
f
and
g
satisfy




11
11
,,
max,< ,
sup sup
mm
st st
fst gst
st st

 






(3)
and satisfy the decay conditions
,bd




1/(1)1/( 1)
00
d<, d<,
pq
ap aq
tbtttdt t



(4)
where , then problem (1) has no positive
entire radial large solution.
=min{, }mpq
In order to state our results conveniently, let us write
 
1/( 1)
11
00
:= ,
lim
=dd
r
p
rt
ap NN
BBr
Brtsbsstr

 
 ,0,
 
1/( 1)
11
00
:= ,
lim
=d
r
q
rt
aq NN
DDr
Drtsds str

 
 d,0
and
  

1/(1)
0
:= ,
lim
d
=,
,,
r
r
m
FFr
s
Fr r
fss gss

>0,
where satisfies
0
m

0
min,, if 1,
=max,, if <1,
pqf g
mpqf g

we see that
  

1/(1)
0
1
=>
,,
m
Fr r
frr grr0,>
so,
F
has the inverse function 1
F
on [, )
, and
F
and 1
F
are both increasing functions on [, )
.
Theorem 2. Assume
=.F
Then the system (1) has infinitely many positive entire
solutions
1
,[0,uv C)
. Moreover, the following
hold:
1) If
<B
and , then u and
are bounded;

<D v
2) If
==BD

, then
=
lim
rr
ur vr
 =
lim  , that is all positive entire
solutions of (1) are large.
Theorem 3. If
Copyright © 2011 SciRes. APM
D. J. LEI ET AL.
Copyright © 2011 SciRes. APM
353
2. Proof of Theorem 1
 
<, <,<,FBD 
and there exist >
and >
such that
 
 
<BDFF ,
  (5)
In this section, we consider the proof of Theorem 1 by
contradictions. Assume that the system (1) has the
positive entire radial large solution . From (1), we
know that
(,)uv
the system (1) has a positive radial bounded solution
satisfying
 
1
,[0,uv C
1/( 1)p
)
;
.
 


1
,
,0
fBrur
FFBrDr r




 
 

1/( 1)
1
,
,0
q
gDrvr
FFBrDr r




Theorem 4. If satisfies
1
m


1
min,, if ,1,
=max,, f ,<1,
pqf g
mpqi fg

then we have
1) If
 
== ,BD 
and
 

1/( 1)
1
,, =0,
lim
m
s
fssgss
s

(6)
then the system (1) has infinitely many positive entire
large solutions;
2) If
 
<, <,BD
and
 

1/( 1)
1
0
,, <
sup m
sfss gss
,
0,
0.
(7)
 

 
1
11
=(,),
p
Nap N
tuttbtfutvtt
 
 

 

1
11
=,,
q
Naq N
tvttdtgutvtt
 
Now we set
 
00
=,=
max max
tr tr
Urut Vrvt
 
,
it is easy to see that are positive and nonde-
creasing functions. Moreover, we have
and
(,)UV
, Uu
Vv
, UrVr
0
C
r
>0
as . It follows from
(3) that there exists such that



1
0
max,,,, for 1,
m
fst gstC stst
 (8)
and
0
max,,,, for 1.fst gstCst (9)
Then by (8) and (9), we have


1
0
max,,,1, for 0.
m
fst gstCstst

(10)
Then we can get
 

 

 

1
0
1
0
,1
1, for 0.
m
m
furvrCurvr
CUrVrr

 
then the system (1) has infinitely many positive entire
bounded solutions. So, for all , we obtain
00rr


 


 


 



 



1/(1)
11
00
0
1/( 1)
1
11
00
0
11/( 1)
11
1
00
0
11/( 1)
1
00
=,d
1dd
1dd
1d
p
rt
ap NN
r
p
rt m
ap NN
r
mp
rt
ap NN
pr
mp
rap
pr
ururtsbs fusvsst
urCtsbsUs Vsst
urCUrVrtsbsst
urCUrVrt btt
 
 
 
 






 



1/(1)
00
1d,
p
rap
r
urCUrVrtbtt

d
q




1/(1)1/( 1)
00
1
maxd ,d<.
4
pq
ap aq
rr
rbrr rdrrC



(11)
where is a positive constant. As , we
have , so the last inequality above is
valid. Notice that (4), we choose such that
C
0<
=min{ ,}mp
>0
1<1mp
0
r
D. J. LEI ET AL.
354
=
0
It follows that , we can
find such that
 
=
lim lim
rr
ur vr

1
rr
 
1
00
= ,=,
maxmax
rtrrtr
UrutVrvtrr
  . (12)
Thus, we have


 



1/(1)
00
1
1
.
p
rap
r
UrurCUr Vrtbtt
rr


d,
By (11), we get


 

0
1, .
4
Ur Vr
Ururr r

 
1
that is
  

11
, .
4
Ur Vr
UrCr r
 
where

10
1
=
4
Cur>0
. Similarly,
 

21
, .
4
Ur Vr
VrCr r
 
which implies
 
12 1
2, UrVrCCr r. (13)
(13) means that U and V are bounded and so
and are bounded which is a contradiction. It follows
that (1) has no positive entire radial large solutions and
the proof is now completed.
u
v
Remark. In Theorem 1, if , and ,>2pq f
g
satisfy

11
,,
max , <
sup sup
st st
fst gst
st st
 







and satisfy the same decay conditions (4), we can
also get the same result that problem (1) has no positive
entire radial large solution.
,bd
In the following,we will give the detailed proof.
Proof. We also consider the proof by contradiction. If
using the same process in Theorem 1, we will omit that
items here.
Assume that the system (1) has the positive entire
radial large solution , we can get from the given
condition above that there exists such that
(,)uv
3>0C

3
max,,,, for 1fst gstCstst

and
3
max,,,, for 1
f
stg stCst
so, we can get

3
max,,,1, for 0.fst gstCstst

Thus, we can get
 
 

3
3
,1
1, for 0,
furvrCurvr
CUrVr r


here
,UrVr are the same functions defined in
Theorem 1.
As the proof of Theorem 1, we omit the same process
here, for all , we obtain
00rr


 



 



 



1/(1)
11
00
0
1/(1)
11
04 0
0
11/(1)
11
1
04 0
0
11/(
1
04 0
=,d
()(1()())dd
1dd
1
p
rt
ap NN
r
p
rt
aq NN
r
p
rt
ap NN
pr
rap
pr
ururtsbs fusvsst
urCts bsUsVs st
urCUrVrtsbsst
urCUrVrt bt
 
 
 
 





1) d
pt
d
where is a positive constant.
4
Notice the condition (4), we choose such that
C
0>0r





1/(1)1/( 1)
00 4
1
maxd ,d<,
pq
ap aq
rr
rbrrrdrrC



together with (12), we get


 

1
1
01p
UrurUr Vr

Similarly,we have


 

1
1
01q
VrvrUr Vr

Set , we get
=min{ , }mpq
 

 

 


 

1
1
00
1
1
1
1
00
1
1
21
p
q
m
Ur VrurvrUr Vr
Ur Vr
urvrUr Vr

 

Copyright © 2011 SciRes. APM
D. J. LEI ET AL.355
that is,
 

 


1
1
00
121
1, 0
m
Ur VrUr Vr
ur vrr
 
 
We claim the above inequality is invalid. In fact, set a
function
 

 

 

1
1
1
:=12 1m
TUrVr
Ur VrUr Vr

 
then
 

 

2
1
2
1=11 >0
1
asis large enough.
m
m
T UrVrUrVr
m
r
 
,
So,
 
1TUrVr
has no positive entire radial large solution and the proof
of the remark is completed.
3. Proofs of Theorem 2 and Theorem 3
Proof of Theorem 2. We start by showing that (1) has
positive radial solutions. On this purpose we fix >
and >
and we show that the system


 



 

 
11
11
=,
=,
0=>0, 0=>0,
,0, on [0,),
Nap N
p
Naq N
q
rurbrfurvrr
rurdrgurvrr
uv
uv

 
 


, >0,
, >0,
(14)
is an increasing function on
,and it can not be always controlled by a fixed
constant, which is a contraction. It follows that system (1)
[0, )
has positive solution (, (where )uv

2
=p
p
s
ss
).
Thus

=,=Ux uxvx
d,0,
d,0.
d, 0,
d, 0.
Vx are positive solutions
of (1). Integrating (14) we have
 

1/(1)
11
00
=,d
p
rt
ap NN
urtsbs fusvsstr
 




1/( 1)
11
00
=,d
q
rt
aq NN
vrtsdsgusvsst r



Let and be the sequences of posi- tive continuous functions defined on [0 by
0
{}
nn
u0
{}
nn
v,)


 

 

0
0
1/( 1)
11
100
1/( 1)
11
100
=,
=,
=,d
=,d
p
rt
ap NN
nnn
q
rt
aq NN
nnn
ur
vr
urt sbsfusvsstr
vrt sbsgusvsstr
 
 



(15)
Obviously, for all , we have
0r
 
0101
, , , .
nn
urvru uv v


The monotonicity of
f
and
g
yield
 
1212
, , 0.urur vrvr r
Repeating such arguments we deduce that

11
, , 0,1,
nn nn
urur vrvr rn


and we obtain that sequences and
are nondecreasing on
0
{}
nn
u0
{}
nn
v
[0, )
. Notice


 



  

 

  


1/( 1)1/( 1)
11
10
1/(1)
1111
1/( 1)
1111
=,d,
(,(,
,
pp
r
ap NN
nnnnn
p
nnnn nnnn
p
nnnn
urrsbsfus vssfur vrBr
f
ur vrurvrBrfur vrur vr
gu rv ru rv rBr
 



 
 
and






 

 

 


1/( 1)1/( 1)
11
10
1/( 1)
1111
1/( 1)
1111
=,d,
(,(,
,,
qq
r
aq NN
nnnnn
q
nnnn nnnn
q
nnnn
vrrsdsgusvssgurvrD r
g
urvrurvrDrfu rv ru rv r
gu rv ru rv rDr





 
Copyright © 2011 SciRes. APM
D. J. LEI ET AL.
356
which implies
 

 



1/(1)
0,
,,
nn
m
nnnn nnnn
ur vrBr Dr
furvrurvr gurvrurvr




where has been defined before. And then integrat- ing on we obtain
0
m(0, )r
  

  


 
1/(1)
00d.
,,
rnn
m
nnnn nnnn
ut vttBr Dr
fut vtut vtgut vtutvt



So


 
() ()
1/(1)
0
d,
,,
urvr
nn
mBr Dr
fg

 

that is
 


,0
nn
Fu rv rFBrDrr

.
(16)
It follows from 1
F
is increasing on [0 and (16)
that
, )
 

1,0
nn
ur vrFFBr Drr

.
th By
(17)
It follows from ()Fat F (17),
the sequences {}
n
u {}
n
v bounded and in-
creasing on 0
[0 or arbitrary 0
cThus,
and have subsequences converging uniformly to
and on 0
[0 . By the arbitrariness of 0
c,
we see that is a positive solution of (15), that is,
is an entire positive solution of (1). Notice
=
and
] f
, ]c
)
1()=.
are
>0. ,c
uv
=
{}
n
u
>0
{}
n
vv
)
= ,U
u
(,
0)V
(,UV
(0) (

and (, )(0,)(0,)

 was chosen
arbitrarily, it follows that (1) has infinitely many positive
entire solutions.
1) If and , then
()<B ()<D
  

1<,ur vrF FBD

 
which implies that are the positive entire
bounded solutions of (1).
(,)UV
2) If , since
()== ()BD
 

1/(1)
1/( 1)
,,
,, 0.
p
q
urfBr vr
gDrr




Thus, we have

==
lim lim
rr
ur vr
 ,
which yield (, are the positive entire large
solutions of (1). The proof of theorem is now completed.
)UV
Proof of Theorem 3. If condition (5) holds, then we
have

 
<.
nn
F
ur vrFBrDr
FBDF




Since 1
F
is strictly increasing on , we have [0,)


1<.
nn
ur vr
FFB D

 
The last part of the proof is clear from the proof of
Theorem 2. The proof of Theorem 3 is now finished.
4. Proof of Theorem 4
1) It follows from the proof of Theorem 3, we have


  


1/(1)
1
1/( 1)
,
,,
p
nn nn
pnnnn
ur urfurvrBr
f
ur vrurvrBr

  (18)
and


  


1/( 1)
1
1/( 1)
,
,.
q
nn nn
qnnnn
vr vrgurvrDr
g
ur vrur vrDr

  (19)
Let be arbitrary. From (18) and (19) we get
>0R
 
  


  

  

 

1/( 1)
1/( 1)
1/( 1)
1
,
,
,
,,
p
nn nnnn
qnnnn
nnnn
m
nnnn
uRvRf uRvRuRvRBR
g uRvRuRvRDR
fuRvRuRvR
gu Rv Ru Rv RBRDRn


 

 
 
1.
Copyright © 2011 SciRes. APM
D. J. LEI ET AL.357
This implies
 
 

 
 
 

1/( 1)
1
[, ,
1
, 1.
m
nnnnnnnn
nn nn
fuRvRuRvR guRvRuRvR
uR vRuR vR
BR DRn




 
]
Taking into account the monotonicity of
, there exists
 
1
nn
n
uRvR



:= .
lim nn
n
LRuRvR

We claim that is finite. Indeed, if not, we let
and the assumption (6) leads us to a con-
tradiction,thus

LR
n
LR is finite. Since are increas-
ing functions, it follows that the map
is nondecreasing and
,
nn
uv
:(0,L)(0, )
   

,
0,, 1.
nn nn
ur vruR vRLR
rRn
 
 
Thus the sequences are bounded from
above on bounded sets. Let
1
{} ,{}
nn nn
uv
1
.
 
:=, :=, for 0.
lim lim
nn
nn
ururvrv rr

Then is a positive solution of (14).
(,)uv
In order to conclude the proof, it is enough to show
that is a large solution of (14). We see
(,)uv


1/(1)
1/( 1)
,,
,, 0
p
q
ur fBr
vrgDr r




Since
f
and
g
are positive functions and
== =BD,
we can conclude that is a large solution of (14)
and so is a positive entire large solution of (1).
Thus any large solution of (14) provide a positive entire
large solution of (1) with
(,)uv
)
(,)UV
(,UV (0) =,(0) =UV
.
Since ( , )(0,)(0,)
 was chosen arbitrarily, it
follows that (1) has infinitely many positive entire large
solutions.
2) If
 

1/( 1)
1
0
,, <
sup m
sfss gss
holds, then we have

:=< .
lim nn
n
LRuRvR
 
Thus


,
0,, 1.
nn nn
ur vruR vRLR
rRn
 
 
So the sequences 1
{} are bounded from
above on bounded sets. Let
1
,{}
nn nn
uv

:=, :=, for 0.
lim lim
nn
nn
ururvrv rr

Then is a positive solution of (14).
(,)uv
It follows from (18) and (19) that is bounded,
which implies that (1) has infinitely many positive entire
bounded solutions. The proof is completed.
(,)uv
5. The Existence and Nonexistence of Entire
Positive Solutions of the Corresponding
Singular Elliptic Systems with Gradient
Term
In this section, we consider the following singular elliptic
systems with gradient term:






21
21
=,, ,
=,, .
ap pp
N
aq qqN
divxuuubxfu vx
divx v v vdxguvx
 
 
 
 
R
R (20)
where , and are continuous, posi-
tive and nondecreasing functions in
3, > 0Na
b d
N
R,
are positive,nondecreasing
and continuous functions.
, :)fg [0,) [0,) [0,
We can get the same four theorems under the same
conditions in the foregoing items. In the detailed proofs,
only a few modifications should be noticed. Such as, we
note
 
1/(1)
11
00
:= ,
lim
=edd, 0
r
p
rt
tapNN
BBr
Brtsbs str


 ,
 
1/( 1)
11
00
:= ,
lim
=edd, 0
r
q
rt
taqN N
DBr
Drtsds str


 ,
Copyright © 2011 SciRes. APM
D. J. LEI ET AL.
358
and
 
  

1/(1)
0
:= ,
lim
d
=,
,,
r
r
m
FFr
s
Fr r
fss gss

>0,
where 0 is defined as before and other changes are
similar, so we omit here.
m
6. Acknowledgements
The authors are grateful to the editor and anonymous
referees for their constructive comments and suggestions,
which led to improvement of the present paper.
7. References
[1] S. Chen and G. Lu, “Existence and Nonexistence of Posi-
tive Solutions for a Class of Semilinear Elliptic Systems,”
Nonlinear Analysis: Theory, Methods & Applications,
Vol. 38, No. 7, 1999, pp. 919-932.
doi:10.1016/S0362-546X(98)00143-6
[2] F. Cìrstea and V. Rădulescu, “Entire Solutions Blowing
up at Infinity for Semilinear Elliptic Systems,” Journal de
Mathématiques Pures et Appliquées, Vol. 81, No. 9, 2002,
pp. 827-846. doi:10.1016/S0021-7824(02)01265-5
[3] C. Yarur, “Existence of Continuous and Singular Ground
States for Semilinear Elliptic Systems,” Electronic Jour-
nal of Differential Equations, Vol. 1, 1998, pp. 1-27.
[4] R. Dalmasso, “Existence and Uniqueness of Positive
Solutions of Semilinear Elliptic Systems,” Nonlinear
Analysis: Theory, Methods & Applications, Vol. 39, No. 5,
2000, pp. 559-568. doi:10.1016/S0362-546X(98)00221-1
[5] A. V. Lair and A. W. Wood, “Existence of Entire Large
Positive Solutions of Semilinear Elliptic Systems,” Jour-
nal of Differential Equations, Vol. 164, No. 2, 2000, pp.
380-394. doi:10.1006/jdeq.2000.3768
[6] Z. D. Yang, “Existence of Entire Explosive Positive Ra-
dial Solutions for a Class of Quasilinear Elliptic Sys-
tems,” Journal of Mathematical Analysis and Applica-
tions, Vol. 288, No. 2, 2003, pp. 768-783.
doi:10.1016/j.jmaa.2003.09.027
[7] H. X. Qing and Z. D. Yang, “A Remark on the Existence
of Entire Positive Solutions for a Class of Quasilinear El-
liptic Problem,” Oriental Journal of Mathematics, Vol. 3,
No. 1, 2010, pp. 43-50.
[8] C. S. Chen, Z. Q. Wang and F. P. Wang,” Existence and
Nonexistence of Positive Solutions for Singular
p-Laplacian Equation in Rn,” Boundary Value Problems,
Vol. 2010, 2010, Article ID 607453, pp. 1-17.
[9] X. G. Zhang and L. S. Liu, “The Existence and Nonexis-
tence of Entire Positive Solutions of Semilinear Elliptic
Systems with Gradient Term,” Journal of Mathematical
Analysis and Applications, Vol. 371, No. 1, 2010, pp.
300-308. doi:10.1016/j.jmaa.2010.05.029
Copyright © 2011 SciRes. APM