AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL

SHELLS AND ITS APPLICATION TO NONLINEAR POISSON

SYSTEMS
arXiv:2411.08713v1 [math.AP] 13 Nov 2024

GENNARO INFANTE, GIOVANNI MASCALI, AND JORGE RODRÍGUEZ–LÓPEZ

Abstract. We provide a new existence result for abstract nonlinear operator systems
in normed spaces, by means of topological methods. The solution is located within the
product of annular regions and conical shells. The theoretical result possesses a wide range
of applicability, which, for concreteness, we illustrate in the context of systems of nonlinear
Poisson equations subject to homogeneous Dirichlet boundary conditions. For the latter
problem we obtain existence and localization of solutions having all components nontrivial.
This is also illustrated with an explicit example in which we also furnish a numerically
approximated solution, consistent with the theoretical results.

1. Introduction

It is well known that the solvability of elliptic systems plays a key role when modelling
real world phenomena [24]. Under the point of view of applications it is of interest to
obtain the explicit solution (if possible) or at least as much qualitative information about
the solution as possible, which can be useful also for devising suitable numerical schemes.
Various methods can be used to provide existence and localization of solutions. Insofar
as topological methods are concerned, a classical approach is to rewrite the differential
problem as an operator system and the localization of the solution of the latter system
yields qualitative informations on the solution of the differential problem. To fix ideas, let
us consider the following system of nonlinear Poisson equations subject to homogeneous

2020 Mathematics Subject Classification. 47H10, 47H11, 45G15, 35J57.

Key words and phrases. Fixed point index, fixed point theorem, nonlinear system, quasilinear elliptic
systems.
1
2 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

Dirichlet boundary conditions


 −∆u = f (x, u, v), in Ω,

−∆v = g(x, u, v), in Ω, (1.1)

u = v = 0, on ∂ Ω,


where Ω ⊂ Rn denotes the unit open ball and f and g are continuous functions. This kind
of systems has been widely investigated by means of different methodologies, for example
variational or topological, see for example the papers [1, 4–7, 11, 15, 17–19, 23, 30, 31], the
reviews [2, 12, 29], and references therein.
When using topological methods to solve the system (1.1), if both the nonlinearities
occurring in (1.1) are sign-changing, it is natural to seek solutions located in the product of
two balls in suitable Banach spaces, while when the nonlinearities f and g are nonnegative,
a natural choice is to look for solutions within suitable cones of positive functions, see,
for instance, [1, 2, 4, 17, 23]. An interesting case occurs when one of the nonlinearities is
nonnegative and the other is sign changing. Our work aims to provide new results that fit
precisely within this framework. For this purpose, we study the following general abstract
problem regarding the solvability of the operator system
(
u1 = T1 (u1 , u2 ),
(1.2)
u2 = T2 (u1 , u2 ).

The main abstract results are based on new fixed point index computations, which pro-
vide sufficient conditions for the existence of coexistence fixed points (u1 , u2 ) for the opera-
tor T = (T1 , T2 ). The term coexistence, already employed by Lan in [19], means that both
components of the fixed point, u1 and u2 , are non-trivial. If only one of the components is
non-trivial, we say that the solution is semi-trivial, see for example [8, 20]. In particular,
to localize the solutions of the system (1.2), we use compression-expansion homotopy type
conditions in each component of the operator T , as in the vector version of Krasnosel’skiı̆
fixed point theorem due to Precup [25,26]. Compared with the original theorem by Precup,
our main result (see Theorem 2.1 below) applies for more general domains of the operator
T and, moreover, we obtain that its fixed point index is either 1 or −1, which in particular
ensures the existence of at least one fixed point.
 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 3

Similar computations of the fixed point index have been already deduced in [16, 28],
also in relation with Krasnosel’skiı̆-Precup fixed point theorem. Unlike [16, 28], here we
work in the context of wedges, instead of restricting the definition of the operators to the
cartesian product of cones, or of cones by closed convex subsets. We underline that this
theoretical generalization has direct consequences in applications, providing new ways to
localize the solutions to (1.2). As far as we are aware, this is the first time that the solution
to (1.2) is located in the cartesian product of an annular region times a conical shell. On
the other hand, the manner how we compute the fixed point index here, which is based
on its multiplicativity property, differs from that in [28] and seems to be more intuitive.
In addition, our results complement previous ones in the literature concerning systems of
nonlinear equations, such as those in [3, 14].
Going back to the applicability of the theoretical results, we consider the existence of
solutions to a Dirichlet system of the form (1.1). We highlight that the nonlinearity g
may be sign-changing. Even so, we provide sufficient conditions for the existence of a
solution (u, v) which is not semi-trivial, positive in u and located within the product of a
conical shell and an annulus. This localization provides interesting qualitative informations
and quantitative estimates on the components of the solution. We also provide a result
useful to construct a numerical approximation of the solution of (1.1). We illustrate in
an explicit example the constants that occur in our theory and we also exhibit numerical
solutions that are consistent with our theoretical approach.

2. Fixed point index computations

For the sake of completeness, we recall first some properties of the fixed point index for
compact maps. Further details can be found in [2] or [13, Chapter 12].
We will say that a closed convex subset K of a normed linear space X is a wedge if λ u ∈ K
for every u ∈ K and for all λ ≥ 0. A wedge K is said to be a cone if K ∩ (−K) = {0}.

Proposition 2.1. Let C be a wedge of a normed space, U ⊂ C be a bounded relatively

open set and S : U → C be a compact map such that S has no fixed points on the boundary
of U (denoted by ∂ U ). Then the fixed point index of S on the set U with respect to C,
iC (S, U ), has the following properties:
4 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

(1) (Additivity) Let U be the disjoint union of two open sets U1 and U2 . If 0 ̸∈ (I −
S)(U \ (U1 ∪ U2 )), then

iC (S, U ) = iC (S, U1 ) + iC (S, U2 ).

(2) (Existence) If iC (S, U ) ̸= 0, then there exists u ∈ U such that u = Su.

(3) (Homotopy invariance) If H : U × [0, 1] → C is a compact homotopy and 0 ̸∈
(I − H)(∂ U × [0, 1]), then

iC (H(·, 0), U ) = iC (H(·, 1), U ).

(4) (Normalization) If S is a constant map with S(u) = u0 for every u ∈ U , then

(
1, if u0 ∈ U,
iC (S, U ) =
0, if u0 ̸∈ U .

(5) (Multiplicativity) For j = 1, 2, let Cj be a wedge, Uj ⊂ Cj be a open bounded set

and Sj : Uj → Cj be a compact map fixed point free on the boundary of Uj . Then

iC1 ×C2 (S1 × S2 , U1 × U2 ) = iC1 (S1 , U1 ) · iC2 (S2 , U2 ).

Proposition 2.2. Assume that S satisfies the hypotheses of Proposition 2.1. Let U be a
bounded relatively open subset of C such that 0 ∈ U .
(a) If λu ̸= Su for all u ∈ ∂ U and all λ ≥ 1, then iC (S, U ) = 1.
(b) If there exists w ∈ C with ∥w∥ =
̸ 0 such that u ̸= Su + λw for every λ ≥ 0 and all
u ∈ ∂ U , then iC (S, U ) = 0.

2.1. Star-shaped convex sets. In the remaining part of this Section, let (X, ∥·∥X ) and
(Y, ∥·∥Y ) be normed linear spaces and K1 ⊂ X, K2 ⊂ Y two wedges. For simplicity, both
norms ∥·∥X and ∥·∥Y will be denoted by ∥·∥.
Let us introduce the concept of star-convex set. We refer the reader to the papers [21,22]
for further properties of star-convex sets and a motivation for working with them in the
context of Krasnosel’skiı̆ type compression–expansion fixed point theorems.

Definition 2.1. We say that a set E ⊂ X is a star convex set if

λx ∈ E for all λ ∈ [0, 1] and all x ∈ E.

 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 5

Note that every convex set containing the zero is a star convex set. The reverse is not
true.

Example 2.1. In the normed space of continuous real functions defined in the compact
interval [0, 1], X = C([0, 1]), the set
 
E = u ∈ X : u ≥ 0, min u(t) < r ,
t∈[a,b]

with [a, b] ⊂ [0, 1] and r > 0, is a star convex set.

However, E is not convex: take [a, b] = [0, 1], r = 1/4 and the functions u1 (t) = t,
u2 (t) = 1 − t, t ∈ [0, 1], to check that u1 , u2 ∈ E whereas (u1 + u2 )/2 ∈
/ E.

For each i = 1, 2, let Ui and Vi be bounded and relatively open subsets of Ki such that
(1) 0 ∈ Vi ⊂ V i ⊂ Ui ;
(2) U i \ Vi is a retract of U i ;
(3) Ui and Vi are star-convex sets.
Now, we compute the fixed point index of a compact map defined in the Cartesian
product of the sets U i \ Vi , i = 1, 2, under component-wise compression-expansion type
assumptions.



Theorem 2.1. Assume that T = (T1 , T2 ) : U 1 \ V1 × U 2 \ V2 → K1 × K2 is a compact
map and, for each i ∈ {1, 2}, there exists hi ∈ Ki \ {0} such that either of the following



conditions are fulfilled in U 1 \ V1 × U 2 \ V2 :
(a) Ti (u) + µ hi ̸= ui if ui ∈ ∂ Vi and µ ≥ 0, and Ti (u) ̸= λ ui if ui ∈ ∂ Ui and λ ≥ 1; or
(b) Ti (u) ̸= λ ui if ui ∈ ∂ Vi and λ ≥ 1, and Ti (u) + µ hi ̸= ui if ui ∈ ∂ Ui and µ ≥ 0.



Then the fixed point index of T in K1 × K2 over U1 \ V 1 × U2 \ V 2 is well-defined
and satisfies that
iK1 ×K2 T, U1 \ V 1 × U2 \ V 2 = (−1)k ,

where k ∈ {0, 1, 2} is the number of times that condition (a) is satisfied, i = 1, 2.

In particular, T has at least one fixed point in U1 \ V 1 × U2 \ V 2 .



Proof. First of all, consider the retraction ρ : U 1 × U 2 → U 1 \ V1 × U 2 \ V2 defined as
ρ(u1 , u2 ) := (ρ1 (u1 ), ρ2 (u2 )), where ρ1 is a retraction of U 1 onto U 1 \ V1 and ρ2 , a retraction
of U 2 onto U 2 \ V2 .
6 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

Next, define the following continuous extension of T to the set U 1 × U 2 ,

N = (N1 , N2 ) : U 1 × U 2 → K1 × K2 , N := T ◦ ρ.

The operator N is compact and, moreover, the fact that N = T ◦ ρ together with assump-
tions (a) and (b) imply that for each i ∈ {1, 2} one of the following conditions is fulfilled
in U 1 × U 2 :

(a∗ ) Ni (u) + µ hi ̸= ui if ui ∈ ∂ Vi and µ ≥ 0, and Ni (u) ̸= λ ui if ui ∈ ∂ Ui and λ ≥ 1; or

(b∗ ) Ni (u) ̸= λ ui if ui ∈ ∂ Vi and λ ≥ 1, and Ni (u) + µ hi ̸= ui if ui ∈ ∂ Ui and µ ≥ 0.

Let us denote C := K1 × K2 . Now, for each i ∈ {1, 2}, take Oi ∈ {Ui , Vi } and consider
the homotopy H : O1 × O2 × [0, 1] → K1 × K2 given by

H((u1 , u2 ), t) = (N1 (u1 , t u2 ), N2 (t u1 , u2 )) .

Clearly, H is well-defined since U1 , V1 , U2 and V2 are star convex sets. Moreover, it follows
from assumptions (a∗ ) and (b∗ ) (with µ = 0 and λ = 1, respectively) that the homotopy


is admissible (i.e., u ̸= H(u, t) for all u ∈ ∂ O1 × O2 and all t ∈ [0, 1]) and thus the
homotopy invariance of the fixed point index ensures that

iC (N, O1 × O2 ) = iC (H(·, 1), O1 × O2 ) = iC (H(·, 0), O1 × O2 ).

Hence we have

iC (N, O1 × O2 ) = iC (Ñ , O1 × O2 ),

where Ñ (u1 , u2 ) = (Ñ1 (u1 ), Ñ2 (u2 )) := (N1 (u1 , 0), N2 (0, u2 )). Therefore, the multiplicativ-
ity property of the fixed point index (see [13, Chapter 12]) guarantees that

iC (N, O1 × O2 ) = iK1 (Ñ1 , O1 ) · iK2 (Ñ2 , O2 ). (2.3)

Now, by the additivity property of the fixed point index we deduce that




iC N, U1 \ V 1 × U2 \ V 2 = iC (N, U1 × U2 ) − iC (N, U1 × V2 ) − iC (N, V1 × U2 )

+ iC (N, V1 × V2 ). (2.4)

Let us consider four cases:

 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 7

Case 1: T1 and T2 satisfy condition (a). Then N1 and N2 satisfy condition (a∗ ), so Propo-
sition 2.2 ensures that

iKi (Ñi , Ui ) = 1 and iKi (Ñi , Vi ) = 0 (i = 1, 2).

Hence, by (2.3), we obtain the following computations of the fixed point index

iC (N, U1 × V2 ) = iC (N, V1 × U2 ) = iC (N, V1 × V2 ) = 0 (2.5)

and
iC (N, U1 × U2 ) = 1. (2.6)

By (2.4), (2.5) and (2.6),

iC N, U1 \ V 1 × U2 \ V 2 = 1.



Finally, since N = T on the set U 1 \ V1 × U 2 \ V2 , we deduce







iC T, U1 \ V 1 × U2 \ V 2 = iC N, U1 \ V 1 × U2 \ V 2 = 1.

Case 2: T1 satisfies condition (a) and T2 , hypothesis (b). Then we have

iK1 (Ñ1 , U1 ) = 1 = iK2 (Ñ2 , V2 ) and iK1 (Ñ1 , V1 ) = 0 = iK2 (Ñ2 , U2 ).

By (2.3), we deduce that

iC (N, U1 × U2 ) = iC (N, V1 × U2 ) = iC (N, V1 × V2 ) = 0, iC (N, U1 × V2 ) = 1

and thus it follows from (2.4) that

iC T, U1 \ V 1 × U2 \ V 2 = iC N, U1 \ V 1 × U2 \ V 2 = −1.

Case 3: T1 satisfies condition (b) and T2 , hypothesis (a). It follows in an analogous way to




Case 2 that iC T, U1 \ V 1 × U2 \ V 2 = −1.
Case 4: Condition (b) holds for both T1 and T2 . In this case, we have

iKi (Ñi , Ui ) = 0 and iKi (Ñi , Vi ) = 1 (i = 1, 2).

Hence,

iC (N, U1 × U2 ) = iC (N, V1 × U2 ) = iC (N, U1 × V2 ) = 0, iC (N, V1 × V2 ) = 1

8 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

and so (2.4) implies that

iC T, U1 \ V 1 × U2 \ V 2 = 1.




In conclusion, iC T, U1 \ V 1 × U2 \ V 2 = ±1 and the existence property of the fixed



point index ensures that T has at least one fixed point located in U1 \ V 1 × U2 \ V 2 . ⊔
⊓

2.2. Cartesian product of annular regions and conical shells. Here, let (X, ∥·∥) and
(Y, ∥·∥) be normed linear spaces such that Y is infinite dimensional and K ⊂ X a cone.
The following notations will be useful: for given r, R ∈ R+ := [0, ∞), 0 < r < R, we
define

Kr,R := {u ∈ K : r < ∥u∥ < R} and K r,R := {u ∈ K : r ≤ ∥u∥ ≤ R}.

Moreover, we denote as Ar,R the following annular region in the normed space Y

Ar,R := {v ∈ Y : r < ∥v∥ < R},

that is, Ar,R = BR \ B r where Bτ stands for the open ball of radius τ centered at the origin
and B τ represents its closure. Furthermore, Ar,R := B R \ Br .
As a direct consequence of Theorem 2.1, we establish a result in the line of the vector
version of Krasnosel’skiı̆ fixed point theorem in cones due to Precup [25, 26].

Theorem 2.2. Take αi , βi > 0, with αi ̸= βi , ri := min{αi , βi } and Ri := max{αi , βi }

for i = 1, 2, assume that T = (T1 , T2 ) : K r1 ,R1 × Ar2 ,R2 → K × Y is a compact map and
that there exist h1 ∈ K \ {0} and h2 ∈ Y \ {0} such that for each i ∈ {1, 2} the following
conditions are satisfied in K r1 ,R1 × Ar2 ,R2 :
(a) Ti (u) + µ hi ̸= ui if ∥ui ∥ = βi and µ ≥ 0;
(b) Ti (u) ̸= λ ui if ∥ui ∥ = αi and λ ≥ 1.
Then the fixed point index of T in K × Y over Kr1 ,R1 × Ar2 ,R2 , iK×Y (T, Kr1 ,R1 × Ar2 ,R2 ),
is well-defined and
iK×Y (T, Kr1 ,R1 × Ar2 ,R2 ) = (−1)k ,

where k ∈ {0, 1, 2} is the number of times that the equality αi = Ri is satisfied, i = 1, 2.

In particular, T has at least one fixed point u = (u1 , u2 ) ∈ K ×Y such that ri < ∥ui ∥ < Ri
for i = 1, 2.
 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 9

Proof. In order to apply Theorem 2.1, take the wedges K1 = K, K2 = Y and the relatively
open sets U1 = BR1 ∩ K, V1 = Br1 ∩ K, U2 = BR2 and V2 = Br2 . Note that, for each
i ∈ {1, 2}, we have that Ui and Vi are star convex sets since they are convex and contain
the zero.
On the other hand, the map ρ1 : B R1 ∩ K → K r1 ,R1 defined as

2
 r v + (r1 − ∥v∥) h1 ,

1 if ∥v∥ < r1 ,
ρ1 (v) = ∥v + (r1 − ∥v∥)2 h1 ∥
v, if r1 ≤ ∥v∥ ≤ R1 ,


is a retraction of U 1 = B R1 ∩ K onto U 1 \ V1 = K r1 ,R1 , see [10, Example 3] or [28]. In

addition, notice that Ar2 ,R2 is a retract of B R2 since in any infinite dimensional normed
space ∂ Br2 is a retract of B r2 .
Therefore, the conclusion follows in a straightforward way from Theorem 2.1. ⊔
⊓

Remark 2.1. Under the assumptions of Theorem 2.2, condition (2.3) can be seen as

iC (N, O1 × O2 ) = iK (Ñ1 , O1 ) · deg(I − Ñ2 , O2 ),

since it follows from the definition of the fixed point index by means of the Leray-Schauder
degree (see [2, 13]) that iY (Ñ2 , O2 ) = deg(I − Ñ2 , O2 ).

Remark 2.2. It is an open problem to decide whether the fixed point index computation
iK×Y (T, Kr1 ,R1 × Ar2 ,R2 ) = (−1)k remains valid provided that T is fixed point free on the
boundary of the set Kr1 ,R1 × Ar2 ,R2 and hypotheses (a) and (b) in Theorem 2.2 are weakened
as

(ā) Ti (u) + µ hi ̸= ui if ∥ui ∥ = βi and µ > 0;

(b̄) Ti (u) ̸= λ ui if ∥ui ∥ = αi and λ > 1.

Notice that the previous approach based on the multiplicativity property of the fixed point
index does not work since it is not possible to guarantee that the operators Ñ1 and Ñ2 are
fixed point free on the boundary of the sets O1 and O2 , respectively.
10 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

3. Application to elliptic systems

Consider the following system of quasilinear elliptic equations subject to Dirichlet bound-
ary conditions 
 −∆u = f (x, u, v), in Ω,

−∆v = g(x, u, v), in Ω, (3.7)

u = v = 0, on ∂ Ω,


where Ω ⊂ Rn denotes the unit open ball in Rn , f : Ω×R+ ×R → R+ and g : Ω×R+ ×R → R

are continuous functions.
In the sequel, in order to apply the theory developed in the previous section, we shall work
with the normed space X = Y = C(Ω) endowed with the usual norm ∥u∥∞ = maxx∈Ω |u(x)|
(we will simply denote ∥·∥ = ∥·∥∞ ) and the cone of nonnegative continuous functions, i.e.,
K := {u ∈ C(Ω) : u ≥ 0}.
Now, we consider the following system of Hammerstein integral equations associated
to (3.7),  Z
 u(x) = k(x, y) f (y, u(y), v(y)) dy,


Ω
Z (3.8)
 v(x) = k(x, y) g(y, u(y), v(y)) dy,


Ω
where k is the Green’s function corresponding to the problem

−∆u = h(x) in Ω, u = 0 on ∂ Ω,

and h is a given continuous function. To the system (3.8) we associate the operator

T = (T1 , T2 ) : K × Y → K × Y,

where Z
T1 (u, v)(x) = k(x, y) f (y, u(y), v(y)) dy,
ZΩ (3.9)
T2 (u, v)(x) = k(x, y) g(y, u(y), v(y)) dy.
Ω
Note that T is well-defined (f ≥ 0 implies that T1 (K × Y ) ⊂ K). Moreover, by the
continuity of f and g, it follows that the operator T is completely continuous.
By a (weak) solution of (3.7), we mean a fixed point of the operator T . Hence, in what
follows, we will apply Theorem 2.2 to the operator T in order to obtain a solution (u, v)
with both components non-trivial. Note that since g is a sign-changing nonlinearity, it is
 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 11

not expected the second component of the solution, v, to be a nonnegative function, but it
will be localized in an annular region and so it cannot be the identically zero function.

Theorem 3.1. Assume that there exists positive numbers 0 < r1 < R1 , 0 < r2 < R2 and
continuous functions f , f , g, g : Ω → R+ such that the following conditions hold:

a) f (x, u, v) ≤ f (x) on Ω × [0, R1 ] × [−R2 , R2 ] and

Z
sup k(x, y)f (y) dy < R1 ;
x∈Ω Ω

b) f (x) ≤ f (x, u, v) on Ω × [0, r1 ] × [−R2 , R2 ] and

Z
sup k(x, y)f (y) dy > r1 ;
x∈Ω Ω

c) |g(x, u, v)| ≤ g(x) on Ω × [0, R1 ] × [−R2 , R2 ] and

Z
sup k(x, y)g(y) dy < R2 ;
x∈Ω Ω

d) g(x, u, v) ≥ 0 on Ω × [0, R1 ] × [−r2 , r2 ], g(x, u, v) ≥ g(x) on Ω × [0, R1 ] × [0, r2 ] and

Z
sup k(x, y)g(y) dy > r2 .
x∈Ω Ω

Then the system (3.7) has at least one weak solution (u, v) such that u is nonnegative,
r1 < ∥u∥ < R1 and r2 < ∥v∥ < R2 .

Proof. Let us apply Theorem 2.2 to the operator T = (T1 , T2 ) : K r1 ,R1 × Ar2 ,R2 → K × Y
defined as in (3.9).
To do so, let us check first that the following conditions concerning the operator T1 are
satisfied in K r1 ,R1 × Ar2 ,R2 :

1) T1 (u, v) ̸= λ u if ∥u∥ = R1 and λ ≥ 1;

2) T1 (u, v) + µ 1 ̸= u if ∥u∥ = r1 and µ ≥ 0 (where 1 denotes the constant function
equal to one).
12 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

To prove 1), we assume by reductio ad absurdum that there exist (u, v) ∈ K × Y with
∥u∥ = R1 , r2 ≤ ∥v∥ ≤ R2 and λ ≥ 1 such that for all x ∈ Ω we have
Z
λ u(x) = k(x, y) f (y, u(y), v(y)) dy
Ω
Z
≤ k(x, y) f (y) dy,
Ω

and thus, taking the supremum on Ω, it follows from condition a) that λ R1 = λ ∥u∥ < R1 ,
a contradiction.
Now, to show that 2) holds, assume to the contrary that there exist (u, v) ∈ K × Y with
∥u∥ = r1 , r2 ≤ ∥v∥ ≤ R2 and µ ≥ 0 such that T1 (u, v) + µ 1 = u, that is, for every x ∈ Ω
we have
Z
u(x) = k(x, y) f (y, u(y), v(y)) dy + µ 1 .
Ω

Then, by hypothesis b) we obtain that for x ∈ Ω,

Z Z
u(x) ≥ k(x, y) f (y, u(y), v(y)) dy ≥ k(x, y) f (y) dy.
Ω Ω

Hence, passing to the supremum on Ω gives the absurd

Z
r1 = sup u(x) ≥ sup k(x, y) f (y) dy > r1 .
x∈Ω x∈Ω Ω

It remains to prove that the operator T2 satisfies the corresponding conditions in the set
K r1 ,R1 × Ar2 ,R2 , namely,

3) T2 (u, v) ̸= λ v if ∥v∥ = R2 and λ ≥ 1;

4) T2 (u, v) + µ 1 ̸= v if ∥v∥ = r2 and µ ≥ 0.

To prove 3), we proceed in a similar way as in the proof of 1) above; note that, in this case,
we have to take care of the absolute value of v, that is
Z
|λ v(x)| = k(x, y) g(y, u(y), v(y)) dy
Ω
Z Z
≤ k(x, y) |g(y, u(y), v(y))| dy ≤ k(x, y) g(y) dy < R2 ,
Ω Ω

which yields a contradiction.

 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 13

Now let us focus on the point 4). Assume that there exist (u, v) ∈ K r1 ,R1 × Ar2 ,R2 with
∥v∥ = r2 and µ ≥ 0 such that T2 (u, v) + µ 1 = v. Then we have that 0 ≤ u(x) ≤ R1 and
−r2 ≤ |v(x)| ≤ r2 for all x ∈ Ω and thus g(x, u(x), v(x)) ≥ 0 for all x ∈ Ω. It follows that
Z Z
v(x) = k(x, y) g(y, u(y), v(y)) dy + µ 1 ≥ k(x, y) g(y, u(y), v(y)) dy,
Ω Ω

which implies v(x) ≥ 0 on Ω. By condition d), we deduce that

Z
v(x) ≥ k(x, y) g(y) dy
Ω

and again, taking the supremum, we get a contradiction.

Therefore, Theorem 2.2 ensures that the operator T has at least one fixed point in
Kr1 ,R1 × Ar2 ,R2 . ⊔
⊓

We have the following result which is helpful to construct a numerical approximation for
the solutions of the system (3.7).

Theorem 3.2. Under the hypotheses a) and c) of Theorem 3.1 it is possible to construct a
weak solution of the system (3.7).

Proof. Take (u, v) ∈ (K ∩ B R1 ) × B R2 and observe that, due to the hypotheses a) and c),
for every x ∈ Ω we have
Z Z
|T1 u(x)| = k(x, y) f (y, u(y), v(y)) dy ≤ k(x, y) f (y) dy < R1 ,
ZΩ Ω
Z
|T2 v(x)| = k(x, y) g(y, u(y), v(y)) dy ≤ k(x, y) g(y) dy < R2 .
Ω Ω

Therefore T maps (K ∩ B R1 ) × B R2 into itself.

Now, take a couple of functions (u0 , v0 ) ∈ (K ∩ B R1 ) × B R2 , and define the sequence
{(un , vn )} as the unique solutions (which exist by classical elliptic theory, see for exam-
ple [9]) of the following systems

 −∆un = f (x, un−1 , vn−1 ), in Ω,

−∆vn = g(x, un−1 , vn−1 ), in Ω, (3.10)

un = vn = 0, on ∂ Ω,


for n = 1, 2, . . ., which are given by

(un , vn ) = T (un−1 , vn−1 ).

14 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

Note that the sequence {(un , vn )} is contained in (K ∩ B R1 ) × B R2 , because of the

previous observation and, furthermore, given the compactness of the operator T , {(un , vn )}
is contained in a compact subset of (K ∩ B R1 ) × B R2 . Therefore there exists a subsequence
of {(un , vn )} (which we denote in the same way, with abuse of notation) that converges to
a couple (u, v) ∈ (K ∩ B R1 ) × B R2 . For this subsequence we have
 Z
 un (x) = k(x, y) f (y, un−1 (y), vn−1 (y)) dy,


Ω
Z (3.11)
 vn (x) = k(x, y) g(y, un−1 (y), vn−1 (y)) dy.


Ω

By means of the Lebesgue dominated convergence theorem, passing to the limit for n → ∞
in (3.11) we obtain
(u, v) = T (u, v),

that is (u, v) is a weak solution of the system (3.7). ⊔

⊓

In the following example we illustrate the applicability of Theorem 3.1 and, using the iter-
ative process illustrated in Theorem 3.2, we construct a numerical solution with properties
consistent with the theoretical predictions.

Example 3.1. Take the open set Ω = {(x, y) ∈ R2 : x2 + y 2 < 1} and consider the system
 1
 −∆u = (1 + x2 )eu (2 + cos v), in Ω,
5




3 (3.12)
−∆v = (1 + x2 )(1 − v 2 )(2 + sin u), in Ω,


 4

u = v = 0, on ∂ Ω.


Note that conditions a) – d) in Theorem 3.1 can be verified by choosing r1 = 1/21,

√
R1 = 1/2, r2 = 1/6, R2 = 3/2 and the constant functions f ≡ 6 e/5, f ≡ 1/5, g ≡ 45/8
and g ≡ 35/24, as the lower and upper bounds of the nonlinearities
1 3
f ((x, y), u, v) = (1 + x2 )eu (2 + cos v) and g((x, y), u, v) = (1 + x2 )(1 − v 2 )(2 + sin u)
5 4
in the corresponding sets. To check these computations take into account that
Z
1 1
sup k((x, y), (w, z)) 1 d(w, z) = sup (1 − x2 − y 2 ) = ,
(x,y)∈Ω Ω (x,y)∈Ω 4 4
as it can be seen by direct calculation.
 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 15

We now numerically approach the above-written system by using the MATLAB solver
for Poisson problems introduced in [27], suitably modified for treating a nonlinear system
of equations by means of the iterative procedure (3.10). We start with the identically zero
initial guess and after fourteen iterations we obtain a numerical solution within a relative
tolerance of 10−10 in the infinity norm, this is illustrated in Figure 1. We remark that the
infinity norms of u and v are 0.191 and 0.406 respectively, these values are consistent with
the estimates obtained with the theoretical results.

(a) The component u (b) The component v

Figure 1. A numerical solution for Example 3.1

Remark 3.1. Note that in Example 3.1 one has that the operator T maps (K ∩B 1/2 )×B 3/2
into itself, therefore a direct application of the Schauder Theorem would yield a solution
of (3.12), but with a less precise localization.

Acknowledgements

G. Infante is a member of the Gruppo Nazionale per l’Analisi Matematica, la Probabilità

e le loro Applicazioni (GNAMPA) of the Istituto Nazionale di Alta Matematica (INdAM)
and of the UMI Group TAA “Approximation Theory and Applications”. G. Mascali is a
member of the Gruppo Nazionale per la Fisica Matematica (GNFM) of INdAM. G. Infante
and G. Mascali are partially supported by the project POS-CAL.HUB.RIA. G. Infante was
partly funded by the Research project of MUR - Prin 2022 “Nonlinear differential prob-
lems with applications to real phenomena” (Grant Number: 2022ZXZTN2). G. Mascali
acknowledges the support from MUR, Project PRIN “Transport phonema in low dimen-
sional structures: models, simulations and theoretical aspects” CUP E53D23005900006. J.
16 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

Rodrı́guez–López has been partially supported by the VIS Program of the University of
Calabria, by Ministerio de Ciencia y Tecnologı́a (Spain), AEI and Feder, grant PID2020-
113275GB-I00 and by Xunta de Galicia, grant ED431C 2023/12.

References

[1] C. O. Alves and D. G. de Figueiredo, Nonvariational elliptic systems, Discrete Contin. Dyn. Syst., 8
(2002), 289–302.
[2] H. Amann, Fixed point equations and nonlinear eigenvalue problems in ordered Banach spaces, SIAM
Rev., 18 4 (1976), 620–709.
[3] M. Beldziński, M. Galewski and I. Kossowski, On a version of hybrid existence result for a system of
nonlinear equations, Adv. Nonlinear Stud., 23 4 (2023), 16 pp.
[4] X. Cheng and Z. Zhang, Positive solutions for a class of multi-parameter elliptic systems, Nonlinear
Anal. Real World Appl., 14 (2013), 1551–1562.
[5] X. Cheng and C. Zhong, Existence of three nontrivial solutions for an elliptic system, J. Math. Anal.
Appl., 327 (2007), 1420–1430.
[6] J. A. Cid and G. Infante, A non-variational approach to the existence of nonzero positive solutions
for elliptic systems, J. Fixed Point Theory Appl., 19 (2017), 3151–3162.
[7] R. Cui, P. Li, J. Shi and Y. Wang, Existence, uniqueness and stability of positive solutions for a class
of semilinear elliptic systems, Topol. Methods Nonlinear Anal., 42 (2013), 91–104.
[8] E. N. Dancer, Positive of Maps and Applications. In: M. Matzeu, A. Vignoli (eds). Topological Nonlin-
ear Analysis. Progress in Nonlinear Differential Equations and Their Applications, vol 15, Birkhäuser,
Boston (1995).
[9] L. C. Evans, Partial Differential Equations, American Mathematical Society, Providence, Rhode Island
(2010).
[10] G. Feltrin, A note on a fixed point theorem on topological cylinders, Ann. Mat. Pura Appl., 196
(2017), 1441–1458.
[11] D. G. de Figueiredo, J. M. do Ó and B. Ruf, Non-variational elliptic systems in dimension two: a
priori bounds and existence of positive solutions, J. Fixed Point Theory Appl., 4 (2008), 77–96.
[12] D. G. de Figueiredo, Semilinear elliptic systems: existence, multiplicity, symmetry of solutions. In:
Handbook of Differential Equations, Vol. 5: Stationary Partial Differential Equations, M. Chipot (ed.),
Elsevier, (2008), 1–48.
[13] A. Granas and J. Dugundji, Fixed Point Theory, Springer, New York (2003).
[14] G. Infante and M. Maciejewski, Multiple positive solutions of parabolic systems with nonlinear, non-
local initial conditions, J. London Math. Soc., 94 (2016), 859–882.
 AN EXISTENCE RESULT IN ANNULAR REGIONS TIMES CONICAL SHELLS 17

[15] G. Infante, M. Maciejewski and R. Precup, A topological approach to the existence and multiplicity
of positive solutions of (p, q)-Laplacian systems, Dyn. Partial Differ. Equ., 12 (2015), 193–215.
[16] G. Infante, G. Mascali and J. Rodrı́guez-López, A hybrid Krasnosel’skiı̆-Schauder fixed point theorem
for systems, Nonlinear Anal. Real World Appl., 80 (2024), 1–9.
[17] K. Q. Lan, Nonzero positive solutions of systems of elliptic boundary value problems, Proc. Amer.
Math. Soc., 139 (2011), 4343–4349.
[18] K. Q. Lan, Existence of nonzero positive solutions of systems of second order elliptic boundary value
problems J. Appl. Anal. Comput., 1 (2011), 21–31.
[19] K. Q. Lan, Coexistence fixed point theorems in product Banach spaces and applications, Math. Meth.
Appl. Sci., 44 (2021), 3960–3984.
[20] K. Q. Lan and W. Lin, Steady-state solutions of one-dimensional competition models in an unstirred
chemostat via the fixed point index theory, Proc. Royal Soc. Edinburgh A, 151 (2021), 240–264.
[21] C. Lois-Prados and R. Rodrı́guez-López, A generalization of Krasnosel’skiı̆ compression fixed point
theorem by using star convex sets, Proc. Royal Soc. Edinburgh A, 150 1 (2020), 277–303.
[22] C. Lois-Prados, R. Precup and R. Rodrı́guez-López, Krasnosel’skiı̆ type compression-expansion fixed
point theorem for set contractions and star convex sets, J. Fixed Point Theory Appl., 22 No. 63 (2020),
1–20.
[23] R. Ma, R. Chen and Y. Lu, Positive solutions for a class of sublinear elliptic systems, Bound. Value
Probl., 2014:28, (2014), 15 pp.
[24] C.V. Pao, Nonlinear Parabolic and Elliptic Equations, Springer, New York, (2013).
[25] R. Precup, A vector version of Krasnosel’skiı̆’s fixed point theorem in cones and positive periodic
solutions of nonlinear systems, J. Fixed Point Theory Appl. 2 (2007), 141–151.
[26] R. Precup, Componentwise compression-expansion conditions for systems of nonlinear operator equa-
tions and applications, Mathematical models in engineering, biology and medicine, 284–293, AIP Conf.
Proc., 1124, Amer. Inst. Phys., Melville, NY (2009).
[27] A. S. Reimer and A. F. Cheviakov, A Matlab-based finite-difference solver for the Poisson problem
with mixed Dirichlet–Neumann boundary conditions, Computer Physics Communications, 184 (2013),
783–798.
[28] J. Rodrı́guez-López, A fixed point index approach to Krasnosel’skiı̆-Precup fixed point theorem in
cones and applications, Nonlinear Anal., 226 (2023), No. 113138, 1–19.
[29] B. Ruf, Superlinear Elliptic Equations and Systems. In: Handbook of Differential Equations, Vol. 5:
Stationary Partial Differential Equations, M. Chipot (ed.), Elsevier, (2008), 211–276.
[30] Z. Zhang and X. Cheng, Existence of positive solutions for a semilinear elliptic system, Topol. Methods
Nonlinear Anal., 37 (2011), 103–116.
18 G. INFANTE, G. MASCALI, AND J. RODRÍGUEZ–LÓPEZ

[31] H. Zou, A priori estimates for a semilinear elliptic system without variational structure and their
applications, Math. Ann., 323 (2002) 713–735.

Gennaro Infante, Dipartimento di Matematica e Informatica, Università della Cal-

abria, 87036 Arcavacata di Rende, Cosenza, Italy
Email address: gennaro.infante@unical.it

Giovanni Mascali, Dipartimento di Matematica e Informatica, Università della Cal-

abria, 87036 Arcavacata di Rende, Cosenza, Italy
Email address: giovanni.mascali@unical.it

Jorge Rodrı́guez–López, CITMAga & Departamento de Estatı́stica, Análise Matemática

e Optimización, Universidade de Santiago de Compostela, 15782, Facultade de Matemáticas,
Campus Vida, Santiago, Spain
Email address: jorgerodriguez.lopez@usc.es