Charro Colorado Peral

MULTIPLICITY OF SOLUTIONS TO UNIFORMLY ELLIPTIC
FULLY NONLINEAR EQUATIONS WITH CONCAVE-CONVEX

RIGHT HAND SIDE
FERNANDO CHARRO, EDUARDO COLORADO, AND IRENEO PERAL
Abstract. We deal with existence, nonexistence and multiplicity of solutions
to the model problem
(P)
8
<
:
F(u
, D
2
u
) = u
q
+u
r
, in ,
u
> 0, in ,
u
= 0, on ,
where R
n
is a smooth bounded domain, F is a 1-homogeneous fully
nonlinear uniformly elliptic operator, > 0, 0 < q < 1 < r < r and r the
critical exponent in a sense to be made precise.
We set up a general framework for F in which there exists a positive thresh-
old for existence and nonexistence. Moreover a result on multiplicity is ob-
tained for 0 < < .
The main diculty comes from the viscosity setting required for this kind
of operators. We also use some degree-theoretic arguments.
The abstract result is applied to several examples, including Puccis ex-
tremal operators, concave (convex) operators and a class of Isaacs operators.
Dedicated to Luis A. Caarelli in his 60th birthday, with our admiration and
friendship.
1. Introduction
Let be a smooth bounded domain, 0 < q < 1 < r and > 0, and consider
F : R
n
S
n
R satisfying the following structural hypothesis,
(F1) Uniform ellipticity: There exist constants 0 < such that for all
X, Y S
n
with Y 0,
trace(Y ) F(, X + Y ) F(, X) trace(Y )
for every R
n
.
(F2) Homogeneity: F(t, tX) = t F(, X) for all t > 0. We further assume
F(0, 0) = 0.
(F3) Structure condition: There exists > 0 such that, for all X, Y S
n
, and
1
,
2
R
n
,
T
,
(X Y ) [
1

2
[ F(
1
, X) F(
2
, Y ) T
+
,
(X Y ) + [
1

2
[,
2000 Mathematics Subject Classication. 35J60, 35B45, 35B33.
Key words and phrases. Fully nonlinear, Uniformly elliptic, Viscosity solutions, Non-proper,
Multiplicity of solutions, Degree theory, Critical exponent, A-priori estimates.
F.C. and I.P. are partially supported by projects MTM2007-65018, MEC, Spain.
E.C. is partially supported by the MEC Spanish projects MTM2006-09282 and RYC-2007-
04136.
1
2 FERNANDO CHARRO, EDUARDO COLORADO, AND IRENEO PERAL
where T
,
are the extremal Pucci operators dened as
T
+
,
(X) =
i
>0
i
(X)
i
<0
i
(X),
T
,
(X) =
i
>0
i
(X)
i
<0
i
(X),
with
i
(X), i = 1, . . . n, the eigenvalues of X. Indeed,
T
,
(X) = inf
AA
,
trace (AX) , T
+
,
(X) = sup
AA
,
trace (AX)
for /
,
=
_
A S
n
: [[
2
A, ) [[
2
R
n
_
. Notice that the structure
condition (F3) amounts to uniform ellipticity when
1
=
2
.
Remark 1. The results we are going to quote from [10] apply to our framework with
their proofs unchanged, since they are based on the ABP estimate, also available
in our problem by hypotheses (F1) and (F3), see for instance [12].
Assuming (F1), (F2), (F3), in [13] was established that there is a number
(0, ) such that the problem
(1)
_
_
_
F(u
, D
2
u
) = u
q
+ u
r
, in ,
u
> 0, in ,
u
= 0, on ,
with 0 < q < 1 < r has at least one nontrivial viscosity solution for < and
no nontrivial solution for > . Notice that the result in [13] holds without any
restriction on the size of r.
Denition 2. Given F : R
n
S
n
R, dene G(X) = F(0, X). We will say that
the operator G : S
n
R satises a Liouville-type result in R
n
up to s whenever
v 0 is the unique non-negative viscosity solution of
(2) G(D
2
v) = v
r
, in R
n
,
for any 1 < r < s. Finally, we dene the critical exponent for problem (1) as,
r = sups [ s R and G satises a Liouville-type result in R
n
for any 1 < r < s .
Notice that (F1) and (F2) imply that G is uniformly elliptic with constants
0 < < and 1-homogeneous.
In Section 3 we will see that blow-up arguments lead to the following problems
for the operator G(X),
(3) G(D
2
v) = v
r
and 0 v(x) v(0) = 1 in R
n
,
and
(4)
_
_
_
G(D
2
v) = v
r
, in R
n
+
0 v(x) v(0, . . . , 0, s) = 1, in R
n
+
v = 0, on R
n
+
,
for some s > 0 and R
n
+
= (x
1
, . . . , x
n
) R
n
: x
n
> 0. Consequently, we will refer
to G as the blow-up operator hereafter.
Our goal in the present work will be to study the existence of a second nontrivial
solution for every (0, ) provided r < r. The proof involves uniform L
estimates and topological arguments.

We assume the following extra hypotheses on F in order to get the L
estimates.
(F4) G(Q
t
XQ) = G(X) where G(X) = F(0, X) as above and Q O(n) = Q
S
n
: Q Q
t
= Id.
(F5) Problems (3), (4) have no positive solution.
MULTIPLICITY OF SOLUTIONS TO UNIFORMLY ELLIPTIC FULLY NONLINEAR. . . 3
The main result of the paper is the following abstract theorem.
Theorem 3. Consider F : R
n
S
n
R satisfying (F1) (F5), and let 0 < q <
1 < r < r. Then, there exist R, 0 < < such that the problem
_
_
_
F(u
, D
2
u
) = u
q
+ u
r
, in ,
u
> 0, in ,
u
= 0, on ,
(i) has no positive solution for > ,
(ii) has at least one positive solution for = ,
(iii) has at least two positive solutions for every (0, ).
In some sense, these results bring to the fully nonlinear framework the well-
known results on global existence and multiplicity of solutions in[2] and [3, 20]
(and the references therein) for the semilinear and quasilinear setting, respectively.
The proof of Theorem 3 follows some ideas of Ambrosetti et al. in [3] and involves
the mentioned uniform L
a-priori estimates.
These kind of bounds are obtained following the blow-up technique of Gidas-
Spruck in [21] which leads to a contradiction with (F5) (see also [18], where dierent
techniques involving topological and variational arguments are developed to get a
priori uniform L
estimates). As a general fact, it is dicult to verify (F5); in this

direction we have the following results.
Theorem 4. Let v be a nontrivial non-negative bounded (viscosity) solution of
(5)
_
_
_
G(D
2
v) = f(v), in R
n
+
v 0, in R
n
+
v = 0, on R
n
+
,
where f is locally Lipschitz continuous function, f(0) 0 and G : S
n
R is
uniformly elliptic with constants 0 < < and 1-homogeneous. Furthermore,
suppose that
G(Q
t
XQ) = G(X) for Q = (q
ij
)
1i,jn
, a matrix with
q
ij
=
ij
if neither i nor j = n, and q
ij
=
ij
otherwise.
(6)
Then, v is monotonic in the x
n
variable:
v
x
n
> 0 in R
n
+
.
The non-existence of solutions to problem (4) follows from the above theorem
(see Section 4 and [5], [27]).
Concerning (3), the following result for general uniformly elliptic nonlinearities
G is proved in [16].
Theorem 5 (Theorem 4.1 in [16]). Let G : o
n
R be a uniformly elliptic operator
in the sense that there exist constants 0 < such that for all X, Y S
n
with
Y 0,
trace(Y ) G(X +Y ) G(X) trace(Y ).
Assume G(0) = 0 and =

(n 1) + 1 > 2, and let v ((R

n
) be a viscosity
solution of
_
G(D
2
v) v
r
, in R
n
,
v 0, in R
n
.
Then, if 0 < r /( 2), we have v 0.
Notice that, even though the exponent /( 2) could not be maximal for a
precise nonlinearity G, Theorem 5 provides a lower bound for the critical exponent,
valid for the whole class of uniformly elliptic operators. Further restrictions on F
allow one to improve the range of exponents. For example, the maximal range for
the linear equation is known to be
1 < r < 2
1 =
n + 2
n 2
,
where 2
= 2n/(n 2) is the Sobolev exponent (see [21]).

Remark 6. We would like to stress that the monotonicity property in Theorem 4
holds whenever r > 1, hence without restriction on the growth of r. This motivates
the fact that the critical exponent in Denition 2 comes from the Liouville result
for (3) alone.
Remark 7. Hypotheses (F4), (F5) are only used in the proof of the uniform L
estimates, Proposition 12, in order to carry out the blow-up argument.

The paper is organized as follows. Section 2 is devoted to some preliminary
results that will be used in the sequel. Then, in Section 3, the proof of Theorem 3
is given. The blow-up argument is presented there and uniform L
a-priori bounds
are stated under hypotheses (F1) (F5). Next, the existence of a second solution
to (1) is proved using a-priori bounds and theoretical degre arguments.
Section 4 is devoted to prove Theorem 4 following [5] and [27].
Finally, Section 5 is devoted to applications of the abstract framework described
in Sections 3 and 4. The examples considered fulll hypotheses (F1) (F5) and
include the model case where G is a Pucci extremal operator (subsection 5.1) which
include the laplacian as a particular case and then, concave (convex) operators
(subsection 5.2) and a class of Isaacs operators, which are neither concave nor
convex (subsection 5.3). We impose hypothesis (F4) in all the cases except when
G is a Pucci operator, where the condition is built-in.
2. Preliminaries
2.1. Eigenvalues. Under hypotheses (F1), (F2) and (F3), Theorem 8 in [6] holds
(here (
2
regularity of is required). Hence, we know that there exists a principal
eigenvalue
1
for F dened as
1
= sup[ v > 0 in s.t. F(v, D
2
v) v.
in the sense that
1
< and there exists a nontrivial solution (eigenfunction) to
(7)
_
_
F(v, D
2
v) = v, in ,
v > 0, in ,
v = 0, on .
Moreover, by denition of
1
, we know that for every >
1
problem (7) does not
have strictly positive solutions.
Other references for the existence of eigenvalues in the fully nonlinear setting are
[27], [8] and the references therein.
The existence of such a principal eigenvalue and eigenfuntion is used in the proof
of Theorem 3.
2.2. Hopf s Lemma for uniformly elliptic equations. We recall here the Hopf
boundary lemma. An adaptation of the proof in [22, Section 3.2] can be found in
[13]. For further renements, see [25] (see also [4] and [28]).
Proposition 8 (Hopfs Lemma). Let be a bounded domain and u a viscosity
solution of
F(u, D
2
u) 0 in ,
where F satises (F1), (F2), and (F3). In addition, let x
0
satisfy
i) u(x
0
) > u(x) for all x .
ii) satises an interior sphere condition at x
0
.
Then, for every nontangential direction pointing into ,
lim
t0
+
u(x
0
+ t) u(x
0
)
t
< 0.
2.3. A strong comparison principle. In the following result, we provide a strong
comparison principle. The main feature is that, once standard comparison is proved,
the new information is fed back in order to get strict comparison.
Proposition 9. Let R
n
be a bounded domain and consider f, g (() with
f g in , and f > 0 in . Consider F : R
n
S
n
R verifying (F1), (F2) and
(F3). Finally, let u, v (() such that
F(u, D
2
u) f(x), and F(v, D
2
v) g(x), in ,
in the viscosity sense. Assume u v on , then u v in . Furthermore, if f < g
in , we have u < v in .
Proof. 1. We can assume v > 0 in , since adding a constant to both u and v
does not aect the problem. Now, let
v
(x) = (1 + ) v(x).
Indeed, by homogeneity,
(8) F(v
, D
2
v
) (1 + ) g(x), and u v
0 on ,
for small enough. Now, we argue by contradiction. Suppose that there exists
x
0
such that
(u v
)(x
0
) = max
(u v
) > 0.
Then, (8) implies x
0
/ . Now dene
w(x, y) = u(x) v
(y)

2
[x y[
2
and denote (x
, y
) such that w(x
, y
) = max
w(x, y). Such pairs (x
, y
)
satisfy
(1) lim
[x
[
2
= 0.
(2) lim
w(x
, y
) = w( x, x) = max
(u v), whenever ( x, x) is an accumulation

point of (x
, y
).
Properties (1) and (2) are well known and can be found, for example, in Lemma
3.1 in [15].
Hence, we can assume x
, y
x
0
as hereafter without loss of generality.
As a consequence, x
, y
for every large enough; applying the Maximum

Principle for semicontinuous functions (see for instance, [14] and [15]), there exist
two symmetric matrices X
, Y
such that
_
(x
), X
_
J
2+
u(x
), and
_
(x
), Y
_
J
2
v
(y
),
and X
in the sense of matrices. By denition of viscosity sub- and superso-

lutions (see [15]), we have
F
_
(x
), X
_
f(x
),
and
F
_
(x
), Y
_
(1 + )g(y
) (1 + )f(y
).
Then, by degenerate ellipticity, we get
(1 + )f(y
) f(x
) F
_
(x
), Y
_
F
_
(x
), X
_
0.
Letting , by continuity we arrive at
0 < f(x
0
) 0,
a contradiction. Thus, u v
in , and, letting 0, we nd u v in .
2. We assume in the sequel that f < g in . If u ,= v we have to prove u < v in .
Since we have already proved that u v in , suppose to the contrary that there
exists x
0
such that u(x
0
) = v(x
0
), that is, x
0
is a maximum point of u v.
Consequently, x
0
is the only maximum point of u(x) v(x) [x x
0
[
4
.
Consider
w(x, y) = u(x) v(y) [x x
0
[
4

2
[x y[
2
and (x
, y
) such that w(x
, y
) = max
w(x, y) as before. By properties (1) and

(2) above, we have x
, y
x
0
as . As a consequence, x
, y
for every
large enough. Reasoning as above, we can nd two symmetric matrices X
, Y
such that
_
(x
), X
_
J
2+
_
u(x
) [x
x
0
[
4
_
, and
_
(x
), Y
_
J
2
v(y
),
and X
in the sense of matrices. As a consequence,

_
(x
) + 4[x
x
0
[
2
(x
x
0
),
X
+ 4[x
x
0
[
2
Id + 4(x
x
0
) (x
x
0
)
_
J
2+
u(x
),
By denition of viscosity sub- and supersolution (see [15]), we have
g(x
) f(y
) F((x
), Y
)
F
_
(x
) + 4[x
x
0
[
2
(x
x
0
),
X
+ 4[x
x
0
[
2
Id + 4(x
x
0
) (x
x
0
)
_
T
,
(Y
) + O([x
x
0
[
2
) O([x
x
0
[
2
).
as 0. Letting 0 we get 0 < g(x
0
) f(x
0
) 0 by hypothesis, and we are
done.
3. The abstract result. Proof of Theorem 3
For the readers convenience we recall the main result already stated in the
introduction.
Theorem 3. Consider F : R
n
S
n
R satisfying (F1) (F5), and let 0 < q <
1 < r < r. Then, there exists R, 0 < < such that the problem
(9)
_
_
_
F(u
, D
2
u
) = u
q
+ u
r
, in ,
u
> 0, in ,
u
= 0, on ,
(ii) has at least one positive solution for = ,
(iii) has at least two positive solutions for every (0, ).
The proof is divided into several steps, organized as subsections.
3.1. Non-existence for large .
Theorem 10. For large enough, problem (9) has no solution in the viscosity
sense.
Proof. Fix >
1
and consider
0
=
rq
r1
(r 1)
_
(1 q)
1q
(r q)
rq
_
1
r1
.
In order to reach a contradiction, suppose that there exists >
0
such that the
problem (9) has a solution u
. Then, we have
(10) F(u
, D
2
u
) = u
q
+u
r
> u
in ,
in the viscosity sense. In fact, it is enough to demonstrate that
min
tR
+
(t) > where
(t) = t
q1
+t
r1
.
It is easy to check that
d
dt
(t) = 0 t
=
_
(1 q)
(r 1)
_ 1
rq
,
which, indeed, is a minimum. Since
(t) both as t 0 and t , it is a

global minimum. Then,
(t
) =
r1
rq
(r q) (1 q)
q1
rq
(r 1)
r1
rq
>
by our selection of . On the other hand, dene =
1
, where
1
is a solution
of (7). Hopfs Lemma (Proposition 8) implies that there exists > 0 such that
u
. Then,
(11) F(, D
2
) =
1
< u
in ,
by denition of .
By construction, we have 0 < u
, where and u
satisfy (10) and (11).

Hence, we can apply the iteration method to get v, satisfying v u, a viscosity
solution of
F(v, D
2
v) = v.
Hence, v is a positive solution of (7), which is a contradiction with the denition
of
1
.
3.2. Existence of one solution for (9) with 0 < . In [13] it is proved
that there exists a threshold > 0 such that problem (9) has at least one positive
solution for 0 < < and no positive solution for > . Here we extend the
result to the critical value = .
Proposition 11. Let R
n
be a smooth bounded domain, and suppose that
F : R
n
S
n
R satises (F1), (F2) and (F3). Let 0 < q < 1 < r < r and > 0.
Then there exists a number > 0 such that problem (9) has at least one solution
for every (0, ].
Proof. Given < , let u
> 0 be the solution to (9) found in [13]. As is pointed

out in Remark 1, it is possible to follow the arguments in Proposition 4.10 in [10] in
order to get Krylov-Safonov (
estimates. Namely, there exists a positive constant

C such that
|u
|
C
()
C, uniformly in ,
since the right hand side of the equation in (9) is uniformly bounded by Proposition
12,
u
q
+ u
r
|u
|
q
+|u
|
r
C.
Consequently, applying the Arzela-Ascoli Theorem, we can nd a sequence u
with
j
as j which converges uniformly to some u
. Such u
j
are
viscosity solutions to problem (9) with =
j
; hence, we can pass to the limit in
the viscosity sense to nd that u
is a solution to problem (9) with = .

In order to prove that u
> 0, notice that by construction |u
j
|
> c >
0 uniformly in j, so |u
> 0. Then we get u
> 0 from the weak Harnack

inequality (see [10])
Thus, statements (i) and (ii) in Theorem 3 are proved. The rest of this section
is devoted to prove the existence of a second solution in (0, ).
3.3. Existence of a second solution in (0, ).
3.3.1. L
estimates: Blow-up argument. We present rst the blow-up method in

[21] adapted to the viscosity setting. We summarize the arguments in the following
result.
Proposition 12. Let F : R
n
S
n
R satisfy (F1)(F5) and let u be a nontrivial
viscosity solution of problem (9) with 0 < q < 1 < r < r. Then there exists a
constant C > 0 independent of and u such that |u|
L
C.
For the proof, we proceed by contradiction. Since we aim to prove that u(x) C
with C = C(r, ) independent of u, suppose to the contrary that there exists a
sequence u
k
k
of positive solutions to
(12)
_
F(u
k
, D
2
u
k
) = u
q
k
+u
r
k
, in ,
u
k
(x) = 0, on ,
and a sequence of points z
k
k
such that
M
k
= sup
u
k
= u
k
(z
k
) as k .
Without loss of generality we can assume z
k
z as k . There are two
cases to be considered, either z or z .
CASE 1: z . Let 2d = dist(z, ) in the sequel. Now, consider
y =
x z
k
k
, x = z
k
+
k
y,
where
2
r1
k
M
k
= 1.
We can then dene
(13) v
k
(y) =
2
r1
k
u
k
(x).
Lemma 13. For k large enough, the function v
k
(y) dened in (13) is a viscosity
solution of
(14) F
_
y
v
k
(y), D
2
y
v
k
(y)
_
=
2(rq)
r1
k
v
q
k
(y) + v
r
k
(y) in B
d/
k
(0).
Proof. In order to prove that v
k
is a viscosity solution of (14), we treat rst the
subsolution case.
Consider (
2
, y
0
B
d/
k
(0) such that v
k
has a local maximum at y
0
.
Indeed, we can suppose without loss of generality that touches v
k
from above at
y
0
, that is,
(v
k
)(y) (v
k
)(y
0
) = 0,
for all y in a neighborhood of y
0
. Dene
(x) =
2
r1
k

_
x z
k
k
_
.
Then, touches u
k
from above at x
0
= z
k
+
k
y
0
(it is here where y
0
B
d/
k
(0)
plays a role), namely
u
k
(x
0
) = u
k
(z
k
+
k
y
0
) =
2
r1
k
v
k
(y
0
) =
2
r1
k
(y
0
) = (x
0
),
and
u
k
(x) = u
k
(z
k
+
k
y) =
2
r1
k
v
k
(y)
2
r1
k
(y) = (x),
for all x in a neighborhood of x
0
.
We can compute the derivatives of (x) in terms of those of (y)
x
(x
0
) =
r1
r1
k

y
(y
0
),
D
2
x
(x
0
) =
2r
r1
k
D
2
y
(y
0
).
Since u
k
is a viscosity subsolution of (12), by homogeneity, we get
F
_
y
(y
0
), D
2
y
(y
0
)
_

2(rq)
r1
k
v
q
k
(y
0
) + v
r
k
(y
0
),
which is what we aimed for. The supersolution case is analogous.
We can x R > 0 and suppose without loss of generality (taking k large enough)
that B
R
(0) B
d/
k
(0).
Our hypotheses on the uniform ellipticity and structure of F imply that
T
,
(D
2
v
k
)
k
[v
k
[
2(rq)
r1
k
v
q
k
(y) + v
r
k
(y), in B
d/
k
(0),
and
T
+
,
(D
2
v
k
) +
k
[v
k
[
2(rq)
r1
k
v
q
k
(y) + v
r
k
(y), in B
d/
k
(0).
In particular, since |v
k
|
L
(B
R
)
= 1 by construction, and
k
0 as k , we
can x > 0 and nd for k large enough that
T
,
(D
2
v
k
) [v
k
[ 1 + , in B
d/
k
(0),
and
T
+
,
(D
2
v
k
) + [v
k
[ (1 + ), in B
d/
k
(0).
Hence, from the Harnack inequality (which follows from the ABP estimate, available
by (F3), see [10] and [12]), we get uniform (
estimates (see [10]),

(15) |v
k
|
C
(B
R/2
)
C(n, R, , , 1 + ),
for some 0 < < 1.
Then, we apply the Arzela-Ascoli Theorem and conclude that there exists a
subsequence v
k
j
and a limit function v such that
lim
k
j
v
k
j
= v, uniformly in B
R
(0), and v(0) = 1.
Indeed, we can consider any R
1
> R and apply the arguments above to the subse-
quence v
k
j
in B
R
1
(0). Then, we get a new subsequence v
k
j
1
such that
lim
k
j
1
v
k
j
1
= v, uniformly in B
R
1
(0), and v(0) = 1.
Notice that, since v
k
j
1
k
j
1
v
k
j
k
j
the limits of both subsequences coincide in
B
R
(0).
We can consider an increasing sequence of radii R
j
j
and iterate this procedure
to get a diagonal subsequence v
k
such that
(16) lim
k
v
k
= v, uniformly in B
R
(0) R > 0, and v(0) = 1.
Finally, we take limits in the viscosity sense in (14), which is the content of the
following lemma.
Lemma 14. The limit v(y) in (16) is a viscosity solution of
_
G
_
D
2
y
v(y)
_
= v
r
(y), in R
n
,
0 v(y) v(0) = 1, in R
n
,
where G(X) = F(0, X).
Proof. Consider (
2
and y
0
such that v has a strict local maximum at y
0
,
that is,
(v )(y) < (v )(y
0
),
for all y ,= y
0
in a neighborhood of y
0
.
Fix R > 0 such that y
0
B
R
(0). By uniform convergence in compact sets, we
deduce that there exist a sequence of points y
k
y
0
as k such that v
k

has a local maximum at y
k
, that is,
(v
k
)(y) (v
k
)(y
k
),
for all y ,= y
k
near y
k
. Without loss of generality, we can further suppose y
k

B
R
(0) B
d/
k
(0) for every k > k
0
.
Then, since v
k
is a viscosity solution of (14), we have
(17) F
_
k
(y
k
), D
2
(y
k
)
_

2(rq)
r1
k
v
q
k
(y
k
) + v
r
k
(y
k
).
Letting k in (17) we arrive at
F
_
0, D
2
(y
0
)
_
v
r
(y
0
),
and we have proved that v is a viscosity subsolution. The supersolution case is
symmetric.
We conclude the argument in this case pointing out that the statement in Lemma
14 contradicts our assumption (F5).
CASE 2: z . In this case the reduction argument leads to a problem in either
R
n
or a half space R
n
+
. Without loss of generality, we can assume that z = 0. In
this way, the tangent space to at the origin is given by x, ) = 0, for some xed
R
n
. Moreover, after a rotation, we can suppose that = (0, ..., 0, 1). Consider
k
dened as in Case 1, see (13), and dene the scaled function
v
k
(y) =
2
r1
k
u(z
k
+
k
y
, z
(n)
k
+
k
y
(n)
)
where z
k
= (z
k
, z
(n)
k
), y = (y
, y
(n)
), with z
k
, y
R
n1
and z
(n)
k
, y
(n)
R.
Consider d
k
=
z
(n)
k
/
k
+o(1) as k , which correspond to the distance from

the maximum of v
k
to the boundary of
k
=
1
k
( (z
k
, z
(n)
k
)).
Then, we have to consider the following alternatives depending on the behavior
of d
k
:
(1) d
k
is unbounded. In this case, passing to the limit in a similar way to
the Case 1, we arrive to the equation G
_
D
2
y
v(y)
_
= v
r
(y) in R
n
, with
0 v(y) v(0) = 1, and we reach a contradiction.
(2) d
k
is bounded. We can take a subsequence, if necessary, such that d
k

s 0.
In the second case there are two alternatives to be considered. If s = 0 we get
a contradiction with the continuity of the limit function v, since on the one hand
v(0) = 1, and on the other hand v(y) = 0 for any y R
n
verifying y, ) = 0, in
particular for y = 0.
If s > 0, we reach the problem
(18)
_
_
_
G(D
2
y
v(y)) = v
r
(y), v 0, in R
n
+
,
0 v(y) v(0, ..., 0, s) = 1, in R
n
+
,
v(x
, 0) = 0, x
R
n1
.
Then, by construction, v in (18) has a maximum at (0, . . . , 0, s). This implies
v(0, . . . , 0, s) = 0 and in particular
v
x
n
(0, . . . , 0, s) = 0,
which contradicts our assumption (F5) and concludes the proof of Proposition 12.
3.3.2. Theoretical degree arguments. Once L
estimates have been proved, we pro-

ceed to the proof of the existence of a second solution using degree theoretic argu-
ments.
Fix (0, ) and consider 0 <
m
< <
M
< . Dene v
M
, w
m
to be
viscosity solutions to
(19)
_
_
_
F(v
M
, D
2
v
M
) =
M
in ,
v
M
> 0 in ,
v
M
= 0 on .
_
_
_
F(w
m
, D
2
w
m
) =
m
d(x), in ,
w
m
> 0, in ,
w
m
= 0, on ,
respectively, where d(x) is the normalized distance to the boundary, that is,
d(x) =
dist(x, )
|dist(, )|
.
It is easy to check (see [13]) that, for t > 0 suciently small, the function
u = t w
m
is a viscosity solution to
(20)
_
_
_
F(u, D
2
u)
m
u
q
+u
r
, in ,
u > 0, in ,
u = 0, on .
In addition, there exists T(
M
) > 0 such that
u = T(
M
) v
M
,
is a viscosity solution to
_
_
_
F(u, D
2
u)
M
u
q
+u
r
, in ,
u > 0, in ,
u = 0, on .
We can assume without loss of generality that t < T(
M
). Indeed, since in the
viscosity sense,
F(u, D
2
u) t
m
< T(
M
)
M
= F(u, D
2
u),
we can apply Proposition 9 in order to get
u 0 in
endowed with the (
1
topology, and
K
(v) := F
1
(f
(v)),
where f
(v) = v
q
+v
r
for simplicity.
The main properties of K
are the following:

(1) K
: X X (see [10] and Remark 1).

(2) u < K
(u) and u > K
(u) in .
Moreover if v K
(u), we have
F(u, D
2
u)
m
u
q
+ u
r
 K
(u), follows in a similar way.

(3) K
is compact in (
1
(see [10] and Remark 1).
Now, dene X as
= v X : u v u.
Indeed, K
: . To see this, let v , that is, v X such that u v u. We

are going to show that u < K
(v) < u. Let w = K
(v); then,
F(u, D
2
u)
m
u
q
+ u
r
< v
q
+ v
r
= F(w, D
2
w).
Again, Proposition 9 implies u < w in and hence u < K
(v). The other inequality

follows analogously.
By the (
1,
estimates in [10] (see Remark 1) and the above computations, we see
that K
is compact. Moreover K
() is a compact set in X. Thus, the Schauder

xed point theorem implies that there exists u
such that K
(u
) = u
, i.e., a
solution to problem (9) with = .
If u
is not the unique xed point in we are done. Otherwise, by Proposition

9, it is easy to show that u 0 such that u
+ B
1
(0) , where B
1
(0) denotes
the unit ball centered at 0 in X.
Proof. For > 0 suciently small, we dene
:= x [ dist(x, ) < .
Previously we have proved that u 0 such that u(x) < Cdist(x, )
for any x
. By Hopfs Boundary Lemma 8 it is easy to conclude the existence

of a constant c > 0 such that c dist(x, ) < u(x) for any x
(for > 0 possibly

smaller than before, depends on the geometric properties of the domain).
Let x
0
and
x
0
= (x
0
) the outward unitary normal to at x
0
. By the
aforementioned (
1,
estimates, we obtain
(21)
t
0
= t
x
0
> 0 such that u(x
0
t
x
0
) < C dist(x
0
t
x
0
, ), 0 < t < t
0
.
Observe that x
0
is arbitrary, so we can cover by
_
x
B
t
x
(x), for t
x
dened
as in (21). By the compactness of , there exists m N, x
j
and t
j
= t
x
j
,
for j = 1, ..., m, verifying (21) and such that
m
j=1
B
t
j
(x
j
). As a consequence,
there exists

t > 0 such that
(22) u(x t
x
) < Cdist(x t
x
, ), 0 < t <

t min
j=1,...,m
t
j
and x .
Thus, we have proved that u(x) < Cdist(x, ) for any x
t
. It is not dicult
to show that u(x) < Cdist(x, ) for any x
t
with C greater than before if
necessary. These arguments prove that there exists a positive constant C verifying
(23) u(x) < Cdist(x, ), x .
Arguing in a similar way, using the Hopf Boundary Lemma 8, instead of the gradient
estimates, we obtain the existence of a constant c > 0 such that
(24) c dist(x, ) < u(x) x .
Using (23), (24) and (
1,
estimates, one can interpolate the distance times a small
positive constant in the inequalities u < u
< u, in the following sense: there exists

0 < << 1 such that
(25) u(x) + dist(x, ) < u
(x) < u(x) dist(x, ), x .

We prove for example the rst inequality: u(x) + dist(x, ) 0 such that w(x
0
t
x
0
) <
0
dist(x
0
t
x
0
, ) for any 0 < t < t
0
, so by continuity one can take
0
> 0 such
that w(x) <
0
dist(x, ) for all x B
0
(x
0
) . Using that x
0
is arbitrary
together with compactness arguments as before, we conclude w(x) < dist(x, )
for all x for some > 0 small enough. By similar arguments we can prove the
second inequality in (25).
It only remains to prove the claim. To this end, we adapt the proof method of
Theorem 5.3 in [10].
Fix H and H
1
such that H
1
H H . Let us denote
u(x) = u(x) and v(x) = u
(x)
for simplicity. Consider their sup- and inf-convolutions (see for instance chapter 5
in [10]), respectively,
u
(x) = sup
yH
_
u(y) +
1
[y x[
2
_
, for x H,
and
v
(x) = inf
yH
_
v(y) +
1
[y x[
2
_
, for x H.
Indeed, u
, v
are, respectively, viscosity solutions to

(27) F(u
, D
2
u
) f
m
_
u +c
2
+ o()
_
in H
1
and
(28) F(v
, D
2
v
) f
_
v c
2
+o()
_
in H
1
where c = max|u|
, |v|
(recall that u, v (
1,
).
We show the details in the case of u
. Let (
2
, and x H such that (u
)(x) (u
)( x) = 0 for all x B = y H : [y x[ < dist( x
, H), with
x
H such that
u
( x) = sup
yH
_
u(y) +
1
[y x[
2
_
= u( x
) +
1
[ x
x[
2
.
Dene (y) = (y + x x
) +
1
[ x
x[
2
. Then, u has a local maximum
at x
, that is, (u )(y) (u )( x
) = 0. Notice that
(29) [ x
x[ |u|

and consequently x
H for small enough. Since x B, we then have y =

x x + x
H.
Hence, since u satises F(u, D
2
u) f
m
(u) in the viscosity sense, we have by
the denition of viscosity solution
F(( x
), D
2
( x
)) f
m
_
u( x
)
_
and, by the denition of ,
F(( x), D
2
( x)) f
m
_
u( x
)
_
.
Finally, since u (
1,
, its Taylors expansion and (29) yields
u( x
) u( x) +|u|
[ x
x[ + o([ x
x[) u( x) + c
2
+ o().
Then we get
F(( x), D
2
( x)) f
m
_
u( x) + c
2
+o()
_
.
and (27) is proved.
Continuing with the proof of the claim, we aim to prove that for small enough,
(30) T
,
_
D
2
(u
)
_
[(u
)[ 0 in H
1
,
in the viscosity sense. Then, since H
1
is arbitrary, and u
converges
uniformly to u v (see [10]), we can pass to the limit in the viscosity sense in (30)
to get (26).
For > 0 small enough, we can x
0
> 0 suciently small to ensure that for
any 0 < <
0
,
(31) f
m
_
u(x) + c
2
+ o()
_
f
_
v(x) c
2
+ o()
_
, x H
where f
(t) = t
q
+t
r
. Clearly,
0
depends on u, v, u, v and H.
Then, for <
0
we consider a paraboloid P touching u
from above at
x
0
H
1
. More precisely, we consider P verifying
_
(u
) P
_
(x)
_
(u
) P
_
(x
0
) = 0, x B
r
(x
0
),
for r > 0 to be xed. We want to prove T
,
_
D
2
P(x
0
)
_
[P(x
0
)[ 0. To this
end, take > 0 and dene
w(x) = v
(x) u
(x) +P(x) + [x x
0
[
2
r
2
.
The regularity of u, v, imply
(32) [w(x)[ = [w(x) w(x
0
)[ C [x x
0
[
< C r
x B
r
(x
0
)
for some positive constants and C depending on u, v,
1
and P. Hence, since P
is xed, we may assume that r is small enough to fulll
(33) |D
2
P| r <

2
, and C r
<

2
.
as well as B
2r
(x
0
) H.
We have w 0 on B
r
(x
0
) and w(x
0
) < 0. Using (b) in Theorem 5.1 in [10], we
know that for any x B
r
(x
0
) there exists a convex paraboloid P
x
of opening K
which touches w from above at x in B
r
(x), where K is a constant independent of
x.
Dene the convex envelope of w in B
r
(x
0
) as
w
(x) = sup
g
_
g(x) : g w in B
r
(x
0
), g convex in B
r
(x
0
)
_
.
The set w =
w
is usually known as the lower contact set of w. We apply Lemma
3.5 in [10] to w in B
r
(x
0
) to show that if x B
r
(x
0
) w =
w
, then P
x
also
touches
w
from above at x in B
r
(x). Indeed, since w(x
0
) < 0, the aforementioned
lemma yields
(34) 0 <
_
B
r
(x
0
){w=
w
}
det D
2
w
.
By (b) in Theorem 5.1 in [10], we know that there exists A B
r
(x
0
) such that
[B
r
(x
0
) A[ = 0, and u
, v
(and hence w) are pointwise second order dierentiable

in A. In fact, by (c) in Theorem 5.1 in [10], equations (27) and (28) are satised
pointwise in A.
Since
w
is convex and
w
w, we have that D
2
w(x) 0 for x Aw =
w
.
It follows from (34) and [B
r
(x
0
) A[ = 0 that
[w =
w
A[ > 0,
and hence, there is at least one point x
1
w =
w
A. At such a point, we have
F
_
u
(x
1
), D
2
u
(x
1
)
_
f
m
_
u(x
1
) + c
2
+ o()
_
,
F
_
v
(x
1
), D
2
v
(x
1
)
_
f
_
v(x
1
) c
2
+ o()
_
,
and D
2
w(x
1
) 0. We deduce that
F
_
u
(x
1
), D
2
u
(x
1
)
_
=
= F
_
v
(x
1
) w(x
1
) +P(x
1
) + 2(x
1
x
0
), D
2
v
(x
1
) D
2
w(x
1
) +D
2
P + 2I
_
F
_
v
(x
1
) +P(x
1
) + 2(x
1
x
0
), D
2
v
(x
1
) + D
2
P + 2I
_
[w(x
1
)[
F
_
v
(x
1
), D
2
v
(x
1
)
_
[w(x
1
)[
+T
,
(D
2
P) [P(x
1
)[ +T
,
(2I) 2[x
1
x
0
[
F
_
v
(x
1
), D
2
v
(x
1
)
_
[w(x
1
)[ +T
,
(D
2
P) [P(x
1
)[ 2(n + r).
Notice that
[P(x
1
)[ =
P(x
0
) + D
2
P
x
1
x
0
[x
1
x
0
[
[x
1
x
0
[
[P(x
0
)[ +|D
2
P| r
All the above inequalities together with (31), (32) and (33) yield,
T
,
(D
2
P) [P(x
0
)[ f
0
_
u(x
1
) + c
2
+ o()
_
f
_
v(x
1
) c
2
+ o()
_
+ 2(n +r) + |D
2
P| r + Cr
2(n +r).
Letting 0, we complete the proof of the claim and hence the Lemma 15.
To complete the proof of the existence of a second xed point in , we follow
the arguments developed in [1] and, more precisely, in [3].
By the permanence and excision properties of degree, we have
(35) deg(I K
, u
+ B
1
(0), 0) = i(K
, u
+B
1
(0), ) = i(K
, , ) = 1.
On the other hand, we recall that problem (9) does not have any positive solution
for > . Moreover, since Proposition 12 provides uniform L
estimates, the
results in [10] yield uniform (
1,
estimates, see also Remark 1. In particular, there
exists C > 0 independent of such that, every u > 0 solution to problem (9)
satises
|u|
C
1 C.
Take > C. Clearly, there are no solution u to problem (9) with |u|
C
1 = . By
the homotopy invariance of the Leray-Schauder degree, we get
deg(I K
, B
1
(0), 0) = deg(I K
+
, B
1
(0), 0) = 0.
Now, by the excision property and (35) we deduce
deg
_
I K
, B
1
(0) u
+ B
1
(0), 0
_
= deg
_
I K
, B
1
(0), 0
_
deg
_
I K
, u
+ B
1
(0), 0
_
= 1.
Hence, K
has another xed point u
B
1
(0) u
+ B
1
(0).
It remains to show that the trivial solution u = 0 has degree 0, i.e., deg
_
I
K
, B
1
(0), 0
_
= 0 for any > 0 suciently small.
To this aim, notice that there exists

such that for any > 0 the problem
(36)
_
F(u, D
2
u) =

u
q
+u
r
+ , u > 0, in ,
u = 0, on
does not have any positive solution. Indeed, arguing as in the proof of Theorem
3.12 in [13], for a given > 0 it is easy to nd

> 0 such that
u
q
+u
r
> (
1
+)u,
and hence, obviously,
u
q
+ u
r
+ > (
1
+ )u,
for any > 0. Following the arguments in the proof of Theorem 3.12 in [13], we
get a positive eigenfunction associated to
1
+ , a contradiction with the results
on existence of eigenvalues in [6], already discussed in Subsection 2.1.
It follows that the homotopy
H(, u) = u F
1
_
u
q
+ u
r
+
_
is admissible and hence
deg
_
I K
, B
1
(0), 0
_
= deg
_
H(0, ), B
1
(0), 0
_
= deg
_
H(1, ), B
1
(0), 0
_
= 0
for all > 0. Then, again by homotopy, one nds
deg
_
I K
, B
1
(0), 0
_
= deg
_
I K
, B
1
(0), 0
_
= 0.
As a consequence, there exists > 0 such that u = 0 is the unique nonnegative
solution of (9) in B
1
(0), which yields (iii) and nish the proof of Theorem 3.
4. A monotonicity property. Proof of Theorem 4
Here we provide the proof of Theorem 4, a monotonicity result in the spirit of
Theorem 3.1 in [27] and Corollary 1.3 in [5]. We recall the statement of the Theorem
for the readers convenience.
Theorem 4. Let v be a nontrivial non-negative bounded (viscosity) solution of
(37)
_
_
_
G(D
2
v) = f(v), in R
n
+
v 0, on R
n
+
v = 0, in R
n
+
,
where f is a locally Lipschitz continuous function, f(0) 0 and G : S
n
R
is uniformly elliptic with constants 0 < < and 1-homogeneous. Furthermore,
suppose that
G(Q
t
XQ) = G(X) for Q = (q
ij
)
1i,jn
a matrix with
q
ij
=
ij
if neither i nor j = n and q
ij
=
ij
otherwise.
(38)
Then, v is monotonic in the x
n
variable:
v
x
n
> 0 in R
n
+
.
Before going into the proof, let us recall some well-known general results in the
form needed below (see for instance [9]).
Proposition 16 (Strong Maximum Principle). Let be a regular domain and let
v be a non-negative viscosity solution to T
+
,
(D
2
v) c(x)v in with c(x) L
.
Then, either v vanishes identically in or v(x) > 0 for all x . Moreover, in the
latter case for any x
0
such that v(x
0
) = 0,
limsup
t0
v(x
0
t) v(x
0
)
t
< 0,
where is the outer normal to .
Proposition 17 (Maximum Principle in narrow domains). Suppose that lies
between two parallel hyperplanes at a distance d. If d is small enough, depending
only on bounds for the coecient c, v is a viscosity solution to T
+
,
(D
2
v) c(x)v
and liminf
x
v(x) 0, then we have v 0 in .
We point out that the symmetry condition (38) is needed in the proof of Theorem
4 in order to be able to apply a moving plane method.
Proof of Theorem 4. Let v be a nontrivial solution to (37). By hypothesis, 0 v
M for some constant M. We can rewrite our equation in the form
G(D
2
v) c(x)v f(0) 0, in ,
for
c(x) =
_
_
_
f
_
v(x)
_
f(0)
v(x)
, if v(x) ,= 0,
0, otherwise.
Since v is bounded, and f is locally Lipschitz then c(x) L
. As a consequence,
the strong maximum principle (Proposition 16) yields v > 0.
We are going to use the moving plane method as in [27]. For each , we dene,
as usual,
T
= x R
n
+
: x
n
= ,
= x R
n
+
: 0 < x
n
< ,
and the functions
v
(x) = v(y, 2 x
n
), w
(x) = v
(x) v(x), x = (y, x

n
) R
n1
R
+
,
dened in
.
First, we point out that
G
_
D
2
v
(x)
_
= f
_
v
(x)
_
in
,
in the viscosity sense. Let us consider for example the subsolution case. Take (
2
and x
0
= (y
0
, x
0
n
)
such that v
has a local maximum at x

0
. Dene
(x) = (y, 2 x
n
). It is easy to see that v
has a local maximum at

(y
0
, 2 x
0
n
). Then, D
2
(y, x
n
) = QD
2
(y, 2 x
n
)Q where Q is a matrix with
elements q
ij
=
ij
if neither i nor j = n and q
ij
=
ij
otherwise. Finally, by
denition of v, we get
f
_
v
(y
0
, x
0
)
_
G
_
D
2
(y
0
, 2x
0
n
)
_
= G
_
QD
2
(y
0
, x
0
n
)Q
_
= G
_
D
2
(y
0
, x
0
n
)
_
in
,
which is what we aimed for.
Next, we have to show that w
= v
v satises
(39) T
+
,
_
D
2
w
(x)
_
c
(x)w
(x),
in the viscosity sense, where
c
(x) =
_
_
_
f
_
v
(x)
_
f
_
v(x)
_
v
(x) v(x)
, if v
(x) ,= v(x)
0, otherwise.
Notice that, again since f is Lipschitz, we have c
(x) L
. The proof follows the

ideas in [17].
To this aim, let (
2
such that w
has a local minimum at some point

x
0

. In other words, x
0
is a local maximum of v v
+ . As usual in the
theory of viscosity solutions, introduce for every > 0
(x, y) = v(x) v
(y) + (x)
[x y[
2
2
[x x
0
[
4
.
For small enough,
attains a maximum in
at some point (x
, y
)
B
r
(x
0
) B
r
(x
0
) for some r > 0. Since x
0
is a local strict maximum of
x v(x) v
(x) + (x) [x x
0
[
4
standard results of the theory of viscosity solutions (see [15]) yields x
, y
x
0
and
|x
|
2
2
0 as 0.
In addition, dening (x, y) = (x) +
|xy|
2
2
+ [x x
0
[
4
, the results in [15]
imply that for any given > 0, there exist matrices X, Y S
n
such that
(
x
(x
, y
), X) J
2,+
v(x
)
(
y
(x
, y
), Y ) J
2,
v
(y
),
(40)
and
_
1
+|A|
_
I
_
X 0
0 Y
_
A + A
2
,
where A = D
2
(x
, y
). From there, setting =

2
, it is standard to see that
X Y D
2
(x
) + O(
2
+[x
x
0
[
2
).
By denition (see [15]) of viscosity solutions and (40), we get
G(X) f
_
v(x
)
_
and G(Y ) f
_
v
(y
)
_
,
and subtracting in the previous inequalities we obtain
f
_
v
(y
)
_
f
_
v(x
)
_
G(Y ) G(X)
G
_
X + D
2
(x
) +O(
2
+[x
x
0
[
2
)
_
G(X)
T
+
,
(D
2
(x
)) +O(
2
+[x
x
0
[
2
).
Letting 0, we get (39).
Then, w
0 in
if is small enough, since w
0 on
and hence we can

apply the maximum principle in narrow domains (Proposition 17).
We dene,
= sup : w
0 in
for all < > 0.

Using Hopfs Lemma, we conclude that w
> 0 in
and
v
x
n
=
1
2
w
x
n
> 0 on T
for every 0 <
. If we prove that
= , we have nished.
Suppose to the contrary that
< . We can x
0
small such the maximum
principle holds for G() c
(x) in
+
0

0
.
Claim: There exists
0
(0,
0
] such that for each (0,
0
) we have w
+
0
in
0
.
Once the claim is proven, we can apply the maximum principle in narrow domains
to
T
+
,
_
D
2
w
(x)
_
c
(x)w
(x) in
0
,
with c
as before, and conclude that w
+
0 in
+
, contradicting the maxi-
mality of
.
Hence, it remains to prove the claim. It follows in a similar way to Lemma 3.1
in [27]. We include the details for the readers convenience.
Suppose that the claim were false, that is, that there exist sequences
m
0
and x
(m)
= (y
(m)
, x
(m)
n
)
0
such that
(41) w
+
m
(x
(m)
) < 0.
We can suppose that x
(m)
n
x
0
n
[
0
,

0
] as m .
We dene the functions
v
(m)
(y, x
n
) = v(y +y
(m)
, x
n
)
and respectively
w
(m)
(y, x
n
) = v
(m)
(y, 2 x
n
) v
(m)
(y, x
n
).
Notice that
G
_
D
2
v
(m)
_
= f
_
v
(m)
(x)
_
.
in the viscosity sense. Then, it is standard to show (see for instance Proposition
4.11 in [10]) that there exists a subsequence and a limit v ( such that v
(m)
v
uniformly in compact sets as m and
G
_
D
2
v
_
= f
_
v(x)
_
in the viscosity sense.
By the strong maximum principle (Proposition 16), we have that either v is
strictly positive in R
n
+
or v 0 in R
n
+
.
Suppose rst that v > 0 in R
n
+
. By what we have already shown, we know that
w
(m)
(y, x
n
) = w
(y + y
(m)
, x
n
) > 0 in
for all
. Hence the limit function

w
= lim
m
w
(m)
is non-negative in
for all
.
So we can repeat the moving plane argument for v and get

, where

is to v what
is to v. Since w
satises
T
+
,
_
D
2
w
(x)
_
c
(x) w
(x)
we can apply the strong maximum principle and get, as before, that w
> 0 in
for all

. On the other hand, by continuity and (41), we have w
(0, x
0
n
) = 0
and x
0
n
(0,

0
], a contradiction.
Suppose next that v 0 in R
n
+
. We x the rectangular domains
Q
1
=
_
x R
n
+
: 1 < x
1
< 1, . . . , 1 < x
n1
< 1,
0
< x
n
< 2
+ 1
_
,
Q
2
=
_
x R
n
+
: 2 < x
1
< 2, . . . , 2 < x
n1
< 2,

0
2
< x
n
< 2
+ 2
_
.
Since v
(m)
converges uniformly to zero in Q
2
, we can suppose that v
(m)
1 in Q
2
for m suciently small. We set
m
= v
(m)
(0, x
(m)
n
) and v
(m)
=
v
(m)
m
.
Now, the function v
(m)
satises
(42) G
_
D
2
v
(m)
(x)
_
=
f
_
v
(m)
(x)
_
v
(m)
(x)
v
(m)
(x), x Q
2
.
The Harnack inequality (see Chapter 4 in [10]) implies
sup
Q
1
v
(m)
C
1
inf
Q
1
v
(m)
C
1
.
Next, we recall that w
0 in
, which implies
v
(m)
(y, x
n
) v
(m)
(y, 2
x
n
) C
1
, for (y, x
n
)
.
Thus, |v
(m)
|
L
(Q)
C
1
, where
Q =
_
x R
n
+
: 1 < x
1
< 1, . . . , 1 < x
n1
< 1, 0 < x
n
< 2
+ 1
_
,
hence, our (
estimates yields (up to a subsequence) that v

(m)
v ( uniformly
in compact sets, and v is a viscosity solution of
G
_
D
2
v
_
lv,
where l = lim
t0
f(t)/t. By the strong maximum principle, either v 0 in Q or v > 0
in Q. The rst possibility is excluded since v(0, x
0
n
) = 1.
We introduce the functions
z
(y, x
n
) = v(y, 2 x
n
) v(y, x
n
)
dened in
Q for all
+ 1/2. We have, by continuity,

z
0 and z
(0, x
0
n
) = 0.
Since
T
+
,
_
D
2
z
(x)
_
lz
(x),
the strong maximum principle, implies z
0 in
Q. This contradicts the

fact that v = 0 on x
n
= 0 and v > 0 on x
n
= 2
.
5. Applications of Theorem 3
In Section 3 we have developed an abstract framework for the global multiplicity
result, Theorem 3. To make it more explicit, we have proved Theorem 3 under
hypotheses (F1) (F5).
The present section is devoted to examples in which it is possible to prove Propo-
sition 12, or in other words, examples for which hypotheses (F4), (F5) hold. The
monotonicity property in Section 4 holds in all the subsequent examples.
We treat rst (subsection 5.1) the important example when G is a Pucci extremal
operator, which include the Laplacian as a particular case.
In subsections 5.2, and 5.3 we treat concave (convex) operators and, respectively,
a class of Isaacs operators, which are neither concave nor convex. We will always
assume that the operators considered satisfy hypothesis (F4), which, in fact, is
built-in in the case of Pucci extremal operators.
It is worth comparing here these examples with the results in [2] where problem
(1) for the operator is studied using variational methods.
The results in [2] are optimal, since the maximal range of exponents,
1 < r < 2
1, is known thanks to the results in [21]. However, the arguments

(in particular those leading to the L
estimates) are strongly dependent on the

structure of the Laplacian.
The examples below lack important features of the Laplacian, such as the varia-
tional structure or classical regularity of the solutions (which is not known in some
cases), which make necessary the use of the viscosity framework in Sections 3 and
4.
5.1. Extremal Pucci operators. In the precise context of F involving a Pucci
extremal operator, that is
F(, X) = T
,
(X) + H(),
with H : R
n
R homogeneous and Lipschitz continuous and such that H(0) = 0,
it is clear that F satises (F1)-(F4).
Concerning (F5), the Liouville-type result in R
n
by Cutri-Leoni (Theorem 5)
applies. On the other hand, Quaas and Sirakov, following the ideas in [5], proved
Theorem 4 in this case, with G() = T
,
().
Observe that T
,
admits (
2,
estimates, in the sense that if the function u is a
viscosity solution to the equation T
,
(D
2
u) = g(x) in a ball B
2R
and g (
for
some (0, 1), then u (
2,
and moreover,
|u|
C
2,
(B
R
)
C
_
|u|
L
(B
2R
)
+|g|
C
(B
2R
)
_
for some constant C > 0. With the aim of these (
2,
-estimates the proof of Theorem
4 could be simplied.
The abstract existence result (Theorem 3) holds and reads as follows.
Theorem 18. Consider R
n
, n 3, a smooth bounded domain, and set
r =
(n 1) +
(n 1)
.
Let H : R
n
R be homogeneous and Lipschitz continuous. Then, for 0 < q < 1 <
r r (or 0 < q < 1 < r < if (n 1) ), there exists R, 0 < <
such that the problem
(43)
_
_
_
T
,
(D
2
u) +H(u) = u
q
+ u
r
, in ,
u > 0, in ,
u = 0, on
(ii) has at least one viscosity positive solution for = ,
(iii) has at least two viscosity positive solutions for every (0, ).
As an important consequence of Theorem 4, Quaas and Sirakov ([27]) deduce a
Liouville-type result in R
n
+
for Pucci operators, adapting arguments in [5] to the
nonlinear setting.
Theorem 19 (Theorem 1.5 in [27]). Suppose n 3 and set
r =
(n 2) +
(n 2)
.
Then the problem
(44)
_
T
,
_
D
2
v
_
= v
r
, in R
n
+
v = 0, on R
n
+
does not have a nontrivial non-negative bounded solution, provided 1 < r r (or
1 < r < if (n 2) ).
Notice that, since the critical exponent in Theorem 19 is greater than the corre-
sponding one in Theorem 5, it is the latter one which yields the critical exponent r
in Theorem 18. This fact agrees with the information obtained from Theorem 4 in
the blow-up argument.
Consequently, whenever the range of exponents r in the Liouville result in R
n
is
maximal, so it is in Theorem 18. Indeed, as we mentioned in the introduction, the
maximal range for the Laplacian (the Pucci operator with = = 1) is known to
be
1 < p < 2
1 =
n + 2
n 2
,
where 2
= 2n/(n 2) is the critical Sobolev exponent. Notice that the exponent

r in Theorem 5 is not optimal in this case.
In fact, in the radial case, Felmer and Quaas [19] proved a Liouville type result for
solutions (instead of just supersolutions as in [16]) for a larger range of exponents
1 < r < r
+
. However, an explicit expression for r

+
in terms of , , n is not
known. When = one gets r
+
= 2
1 as in [21] which is, as we have already

mentioned, optimal also in the non-radial case. When < , it is known that
r
+
> max r, 2
1. As far as we know, to establish the Liouville result in the

range r < r < r
+
is an open problem.
5.2. Concave and convex operators. Let F : R
n
S
n
R satisfy (F1)-(F4),
and suppose that the blow-up operator G(X) = F(0, X) is concave (or convex).
The hypothesis on G being concave (or convex), yields the following regularity
result by Evans and Krylov. For the proof we refer to [10], Section 8.1.
Theorem 20. Let G : S
n
R be a uniformly elliptic (with constants 0 < < )
concave (convex) operator. If the function u is a viscosity solution to the equation
(45) G(D
2
u) = g(x)
in a ball B
2R
and g (
for some (0, 1), then u (

2,
and moreover,
|u|
C
2,
(B
R
)
C
_
|u|
L
(B
2R
)
+|g|
C
(B
2R
)
_
.
In addition, if (45) is satised in a regular domain and u = 0 on the boundary of
the domain, then u satises a (
estimate up to the boundary.

The Liouville-type result in R
n
by Cutri-Leoni (Theorem 5) applies in this case.
For the R
n
+
case, we use Theorem 4 which implies that problem (4) does not have a
solution and allows us to reach a contradiction and conclude the blow-up argument.
As in the case of Pucci operators, the classical regularity of solutions (Theorem
20) simplies the proofs.
The arguments in Sections 3 and 4 shows that Theorem 3 reads as follows.
n
, n 3, a smooth domain. Let F : R
n
S
n
R
be an operator satisfying (F1), with constants 0 < , and (F2) to (F4). In
addition, suppose that G(X) = F(0, X) is concave (convex). Finally, set
r =
(n 1) +
(n 1)
.
Then for 0 < q < 1 < r r < (or 0 < q < 1 < r < if (n 1) ), there
exists R, 0 < < such that the problem
_
_
_
F(u, D
2
u) = u
q
+u
r
, in ,
u > 0, in ,
u = 0, on ,
5.3. A class of Isaacs operators. Finally we intend to apply the above analysis
to a class of operators which are neither convex nor concave and for which classical
regularity is not known in general.
Consider the class of Isaacs operators
(46) F(, X) = sup
lL
inf
kK
_
L
k,l
(, X)
_
,
where / and L are arbitrary sets of indexes and L
k,l
are of the form
L
k,l
(, X) = trace(A
k,l
X) + H
k,l
()
for H
k,l
: R
n
R 1-homogeneous and Lipschitz continuous (all with the same
constant), such that H
k,l
(0) = 0 and A
k,l
is a family of matrices with the same
ellipticity constants.
In addition, we assume that the symmetry hypothesis (F4) holds for the operator
G(X) = F(0, X). More precisely, we require that for every Q O
n
, whenever
A
k,l
S
n
gives rise to an operator L
k,l
, with (k, l) / L, the matrix QA
k.l
Q
t
gives rise to L
k,
l
for some other pair (
k,
l) / L.
Classical regularity is not known in general for problems of the form (46). How-
ever in [11] classical regularity is proved for Isaacs operators of the particular form
(47) F(, X) = min
_
F
(, X), F
(, X)
_
, R
n
and X S
n
with F
: R
n
S
n
R and F
: R
n
S
n
R uniformly elliptic operators
respectively convex and concave in the matrix argument. Clearly F in (47) is neither
concave nor convex. As before, a model operator satisfying all the hypotheses above
could be
F(u, D
2
u) = min
_
inf
kK
L
k
u, sup
lL
L
l
u
_
where
L
k
u = trace(A
k
D
2
u) + H
k
(u).
Notice that, since classical regularity is not known in general, the viscosity setting
becomes crucial in this case. Arguing as in Sections 3 and 4 we have the following
result.
n
, n 3, a smooth domain. Let F : R
n
S
n
R be
an operator of the form (46) satisfying (F1), with ellipticity constants 0 < ,
(F2), (F3) and (F4). Finally, set
r =
(n 1) +
(n 1)
.
Then, for 0 < q < 1 < r r (or 0 < q < 1 < r < if (n 1) ), there exists
R, 0 < < such that the problem
_
_
_
F(u, D
2
u) = u
q
+u
r
, in ,
u > 0, in ,
u = 0, on ,
Acknowledgements. The authors want to thank the anonymous referee for his
comments which improved the nal form of the manuscript, and also Boyan Sirakov
for some helpful discussions.
References
[1] H. Amann; Fixed point equations and nonlinear eigenvalue problems in ordered Banach
spaces, SIAM Rev. 18, no. 4 (1976), 620-709.
[2] A. Ambrosetti, H. Brezis, G. Cerami; Combined eects of concave and convex nonlinearities
in some elliptic problems, J. Funct. Anal. 122, no. 2 (1994), 519-543.
[3] A. Ambrosetti, J. Garca Azorero, I. Peral; Multiplicity results for some nonlinear elliptic
equations, J. Funct. Anal. 137 (1996), 29-242.
[4] M. Bardi, F. Da Lio; On the strong maximum principle for fully nonlinear degenerate elliptic
equations, Arch. Math. (Basel) 73 (1999), no. 4, 276-285.
[5] H. Berestycki, L. Caarelli, L. Nirenberg; Further qualitative properties for elliptic equations
in unbounded domains (Dedicated to Ennio De Giorgi), Ann. Scuola Norm. Sup. Pisa Cl.
Sci. (4) 25 (1997), no. 1-2, 69-94.
[6] I. Birindelli, F. Demengel; Eigenvalue, maximum principle and regularity for fully nonlinear
homogeneous operators, Comm. on Pure and Appl. Analysis, 6, (2007) no. 2, 335-366.
[7] L. Boccardo, M. Escobedo, I. Peral; A Dirichlet Problem Involving Critical Exponents, Non-
linear Analysis, Theory, Methods & Applications, 24 (1995), no. 11, 1639-1648.
[8] J. Busca, M. J. Esteban, A. Quaas; Nonlinear eigenvalues and bifurcation problems for
Puccis operators, Ann. I. H. Poincare, 22, (2005) 187-206.
[9] X. Cabre; Nondivergent elliptic equations on manifolds with nonnegative curvature. Comm.
Pure Appl. Math. 50 (1997), no. 7, 623-665.
[10] X. Cabre, L.A. Caarelli; Fully Nonlinear Elliptic Equations, Amer. Math. Soc., Colloquium
publications, Vol. 43 (1995).
[11] X. Cabre, L.A. Caarelli; Interior C
2,
regularity theory for a class of nonconvex fully non-
linear elliptic equation, Journal de Mathematiques Pures et Appliques, 82 (2003), no. 5,
573-612
[12] L.A. Caarelli, M.G. Crandall, M. Kocan, A.

Swiech; On viscosity solutions of fully nonlinear
equations with measurable ingredients, Communications on Pure and Applied Mathematics,
49 (1996), 365-397.
[13] F. Charro, I. Peral; Zero Order Perturbations to Fully Nonlinear equations: Comparison,
existence and uniqueness (to appear in Comm. in Contemp. Math.).
[14] M. G. Crandall, H. Ishii; The Maximum Principle for Semicontinuous Functions, Dierential
and Integral Equations 3 (1990), no. 6, 1001-1014.
[15] M. G. Crandall, H. Ishii, P. L. Lions; Users Guide to Viscosity Solutions of Second Order
Partial Dierential Equations, Bull. Amer. Math. Soc. 27 (1992), no. 1, 1-67.
[16] A. Cutri, F. Leoni; On the Liouville Property for Fully Nonlinear Equations, Ann. Inst. H.
Poincare, Analyse Non Lineaire 17 (2000), no. 2, 29-245.
[17] F. Da Lio, B. Sirakov; Symmetry results for viscosity solutions of fully nonlinear uniformly
elliptic equations (preprint, 2007).
[18] D. G. de Figueiredo, P.L. Lions, R. D. Nussbaum; A priori estimates and existence of positive
solutions of semilinear elliptic equations, J. Math. Pures Appl. 61 (1982), no. 1, 41-63.
[19] P. Felmer, A. Quaas; On critical exponents for the Puccis extremal operators, Ann. Inst. H.
Poincare Anal. Non Lineaire 20 (2003), no. 5, 843-865.
[20] J. P. Garca Azorero, J.J. Manfredi, I. Peral Alonso, Sobolev versus Holder local minimizers
and global multiplicity for some quasilinear elliptic equations. Commun. Contemp. Math. 2
(2000), no. 3, 385-404.
[21] B. Gidas, J. Spruck; A Priori Bounds for Positive Solutions of Nonlinear Elliptic Equations,
Comm. Partial Dierential Equations 6 (1981), no. 8, 883-901.
[22] D. Gilbarg, N. S. Trudinger; Elliptic Partial Dierential Equations of Second Order, Springer-
Verlag, New York (1983).
[23] P. Juutinen; Principal eigenvalue of a badly degenerate operator (preprint, 2006).
[24] P. Juutinen, P. Lindqvist and J. Manfredi, The -eigenvalue problem, Arch. Ration. Mech.
Anal. 148 (1999), no. 2, 89-105.
[25] B. Kawohl, N. Kutev; Strong Maximum Principle for Semicontinuous Viscosity Solutions of
Nonlinear Partial Dierential Equations, Arch. Math. 70 (1998), 470-478.
[26] A. Quaas; Existence of positive solutions to a semilinear equation involving the Puccis
operator in a convex domain, Di. Int. Eq. 17 (2004), 481-494.
[27] A. Quaas, B. Sirakov; Existence Results for Nonproper Elliptic Equations Involving the Pucci
Operator, Comm. in Partial Dierential Equations, 31 (2006), no. 7, 987-1003.
[28] N. Trudinger; Comparison Principles and Pointwise Estimates for Viscosity Solutions of
Nonlinear Elliptic Equations, Revista Matematica Iberoamericana 4 (1988), 453-468.
Departamento de Matem aticas, Univ. Aut onoma de Madrid, 28049 Madrid, Spain.
E-mail address: fernando.charro@uam.es, ireneo.peral@uam.es
Departamento de Matem aticas, Univ. Carlos III de Madrid, 28911 Madrid, Spain.
E-mail address: eduardo.colorado@uc3m.es

Charro Colorado Peral

Transféré par

Informations du document

Description originale:

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Charro Colorado Peral

Transféré par

Droits d'auteur :

Formats disponibles

MULTIPLICITY OF SOLUTIONS TO UNIFORMLY ELLIPTIC

FULLY NONLINEAR EQUATIONS WITH CONCAVE-CONVEX

estimates and topological arguments.

estimates). As a general fact, it is dicult to verify (F5); in this

(n 1) + 1 > 2, and let v ((R

= 2n/(n 2) is the Sobolev exponent (see [21]).

estimates, Proposition 12, in order to carry out the blow-up argument.

) such that w(x

w(x, y). Such pairs (x

(u v), whenever ( x, x) is an accumulation

for every large enough; applying the Maximum

in the sense of matrices. By denition of viscosity sub- and superso-

) such that w(x

w(x, y) as before. By properties (1) and

in the sense of matrices. As a consequence,

(t) > where

(t) both as t 0 and t , it is a

satisfy (10) and (11).

> 0 be the solution to (9) found in [13]. As is pointed

estimates. Namely, there exists a positive constant

is a solution to problem (9) with = .

> 0, notice that by construction |u

> 0. Then we get u

> 0 from the weak Harnack

estimates: Blow-up argument. We present rst the blow-up method in

estimates (see [10]),

+o(1) as k , which correspond to the distance from

estimates have been proved, we pro-

are the following:

: X X (see [10] and Remark 1).

(u) and u > K

(u) in . The second

(u), follows in a similar way.

: . To see this, let v , that is, v X such that u v u. We

(v) < u. Let w = K

(v). The other inequality

() is a compact set in X. Thus, the Schauder

is not the unique xed point in we are done. Otherwise, by Proposition

< u in . Moreover, we have the following result.

< u; then using the (

. By Hopfs Boundary Lemma 8 it is easy to conclude the existence

(for > 0 possibly

< u, in the following sense: there exists

(x) < u(x) dist(x, ), x .

(x) for all

are, respectively, viscosity solutions to

)( x) = 0 for all x B = y H : [y x[ < dist( x

, that is, (u )(y) (u )( x

H for small enough. Since x B, we then have y =

(and hence w) are pointwise second order dierentiable

has another xed point u

(x) v(x), x = (y, x

has a local maximum at x

has a local maximum at

. The proof follows the

has a local minimum at some point

). From there, setting =

if is small enough, since w

and hence we can

for all < > 0.

for every 0 <

as before, and conclude that w

. Hence the limit function