0% found this document useful (0 votes)
8 views

Paper- final version

Uploaded by

Kai To
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
8 views

Paper- final version

Uploaded by

Kai To
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

Received <day> <Month>, <year>; Revised <day> <Month>, <year>; Accepted <day> <Month>, <year>

DOI: xxx/xxxx

RESEARCH ARTICLE

Nehari manifold and bifurcation for a 𝜓-Hilfer fractional


𝑝-Laplacian

J. Vanterler da C. Sousa*

Department of Applied Mathematics,


Imecc-State University of Campinas, Abstract
13083-859, Campinas, SP, Brazil.
In this paper, we discuss the bifurcation from infinity for nonlinear problem with
Correspondence a fractional 𝑝-Laplacian in 𝜓-fractional space ℍ𝑝𝛼,𝛽;𝜓 , whose bifurcation is directly
* J. Vanterler da C. Sousa, 2 Department of 𝑇
Applied Mathematics, Imecc-State
𝐛 (𝑥) ||𝜙1 || dx with 1 < 𝑞 < 𝑝 − 1.
𝑞+1
University of Campinas, 13083-859,
linked to the signal

Campinas, SP, Brazil. Email: 0
[email protected]
KEYWORDS:
Fractional 𝑝-Laplacian, bifurcation from infinity, Nehari Manifold, Fibering Maps.

1 INTRODUCTION

Consider the nonlinear problem with a fractional 𝑝-Laplacian given by


⎧ ( )
𝛼,𝛽;𝜓 |𝐻 𝛼,𝛽;𝜓 |𝑝−2 𝐻 𝛼,𝛽;𝜓
⎪ D𝑇
𝐻
| D0+ 𝜉 (𝑥)| D0+ 𝜉 (𝑥) = 𝜆 |𝜉 (𝑥)|𝑝−2 𝜉 (𝑥) + 𝐛 (𝑥) |𝜉 (𝑥)|𝑞−1 𝜉 (𝑥)
⎨ | | (1)
⎪ 𝐈𝛽(𝛽−1);𝜓 𝜉 (0) = 𝐈𝛽(𝛽−1);𝜓 𝜉 (𝑇 )
⎩ 0+ 𝑇

1
where 𝐻 D𝛼,𝛽;𝜓
0+
(⋅) , 𝐻 D𝛼,𝛽;𝜓
𝑇
(⋅) are 𝜓-Hilfer fractional derivatives left-sided and right-sided of order 𝑝
< 𝛼 < 1, type 0 ≤ 𝛽 ≤ 1,
𝛽(𝛽−1);𝜓
1 < 𝑞 < 𝑝 − 1, 𝑏 ∈ 𝐿 (Ω) and ∞
𝐈0+ (⋅) and 𝐈𝛽(𝛽−1);𝜓
𝑇
are 𝜓-Riemann-Liouville fractional integrals left-sided and
(⋅)
𝛼,𝛽;𝜓
right-sided, for all 𝑥 ∈ Ω = [0, 𝑇 ]. We also that 𝜉 ∈ ℍ𝑝 and 𝜆
> 0.
Fractional variational problems over the decade have been gaining prominence in the academic community. Currently, we can
find many works that discuss important and non-trivial results about existence, non-existence, multiplicity of solutions of the
fractional 𝑝-Laplacian, fractional Dirichlet problems among others 1,2,3,4 , and references therein. In 2020, Sousa et al. 5 , through
the 𝜓-Hilfer fractional operator formulated fractional 𝑝-Laplacian (problems) and discussed the properties of existence and
non-existence of weak solutions to this problem, in addition to other important properties.
In 2017, Ledesma 6 , discussed results of the existence of non-trivial solutions using the mountain pass theorem for a fractional
𝑝-Laplacian problem in ℝ𝑛 . In 2019, Nyamoradi and Tersian 7 , attacked a fractional 𝑝-Laplacian problem via critical point theory
and variational method, in order to obtain a new criterion for the existence of solutions to the attacked problem. Numerous
papers in this sense have been discussed, some of them, we highlight 3,4,8 .
In 2016 Fiscella et. al 1 discuss bifurcation and multiplicity results for the following critical nonlocal fractional Laplacian
problem
⎧ 2∗ −2
⎪ ∫ (𝑢 (𝑥 + 𝑦) + 𝑢 (𝑥 − 𝑦) − 2𝑢 (𝑥)) 𝐾 (𝑦) 𝑑𝑦 = 𝜆𝑢 + |𝑢| 𝑢, in Ω
⎨ ℝ𝑛
⎪ 𝑢 = 0, in ℝ𝑛 ∕Ω

whose model ig given y the fractional Laplacian − (−Δ) 0 < 𝑠 < 1, Ω is an open bounded subset of ℝ𝑛 , 𝑛 > 2𝑠, with continuous
𝑠
2𝑛
boundary, 𝜆 is a positive real parameters, 2∗ ∶= (𝑛−2𝑠) , is the fractional critical Sobolev exponent.
2 J. Vanterler da C. Sousa ET AL

In 2019 Li and Sun 9 , investigate the existence and bifurcation of nontrivial solution for the following critical nonlocal problem
{ ∗
−L𝐾 𝑢 = 𝜆𝑓 (𝑥, 𝑢) + |𝑢|2 −2 𝑢, in Ω
𝑢 = 0, in ℝ𝑛 ∕Ω
2𝑛
where 𝛼 ∈ (0, 1), Ω is an open bounded subset of ℝ𝑛 (𝑛 > 2𝛼) with continuous boundary, 𝜆 is a positive real parameter, 2∗ ∶= 𝑛−𝛼
is the fractional critical Sobolev exponent, 𝑓 satisfies suitable growth condition and L𝐾 is the integrodifferential operator defined
as
L𝐾 𝑢 (𝑥) = (𝑢 (𝑥 + 𝑦) + 𝑢 (𝑥 − 𝑦) − 2𝑢 (𝑥)) 𝐾 (𝑦) 𝑑𝑦, 𝑥 ∈ ℝ𝑛 . (2)

ℝ𝑛
We can highlight is the great attention that has been given to papers related to the existence, multiplicity, bifurcation of
solving problems involving Laplacian, in several directions 9,10,11,12 and reference therein. We highlight here papers in this sense,
which involve classic fractional operators, such as: Riemann-Liouville, Caputo, Weyl 13,14,15 . Recently, the paper of Sousa et
al. 5 , involving the non-local fractional operator, called, 𝜓-Hilfer. This front of work related to the 𝜓-Hilfer fractional operator,
has recently started by a group of researchers: J. Vanterler, E. Capelas, J. Zuo, Donal O’Regan, Leandro Tavares and Cesar
Ledesma 16,17 . What has been noticed, is a wide path and open problems that deserve attention, as discussed in the papers, and
in the work addressed here.
The existence of solutions of the problem (1), is given by the following result:
5
Theorem 1. Suppose that there exists 𝜆̄ > 0 such that for any 𝜆 < 𝜆,
̄ we have 𝐋− (𝜆) ⊆ 𝚼− . Then,

1. There is a minimizing point for 𝐄𝜆 in 𝐍+𝜆 ,

2. There is a minimizing point for 𝐄𝜆 in 𝐍−𝜆 , proving that 𝐋− (𝜆) is non-empty.

The following notations: 𝐋− (𝜆), 𝚼− (𝜆), 𝚼+ (𝜆), 𝐍−𝜆 and 𝐍+𝜆 , will be presented in a later section.
Definition 1. 5 Let 0 < 𝛼 ≤ 1, 0 ≤ 𝛽 ≤ 1 and 1 < 𝑝 < ∞, The 𝜓-fractional derivative space ℍ𝑝𝛼,𝛽;𝜓 ∶= ℍ𝛼,𝛽;𝜓 𝑝 ([0, 𝑇 ] , ℝ) is
defined by the closure of 𝐶0∞ ([0, 𝑇 ] , ℝ) , and is given by
{ }
ℍ𝛼,𝛽;𝜓
𝑝
= 𝜉 ∈ 𝐿𝑝 ([0, 𝑇 ] , ℝ) ; 𝐇 D𝛼,𝛽;𝜓
0+
𝜉 ∈ 𝐿𝑝 ([0, 𝑇 ] , ℝ) , 𝐈𝛽(𝛽−1)
0+
𝜉 (0) = 𝐈𝛽(𝛽−1)
𝑇
𝜉 (𝑇 ) = 0
= 𝐶0∞ ([0, 𝑇 ] , ℝ) (3)
with the following norm
( )
‖ ‖𝑝 1∕𝑝
‖𝜉‖ℍ𝑝𝛼,𝛽;𝜓 = ‖𝜉‖𝑝𝐿𝑝 + ‖𝐇 D𝛼,𝛽;𝜓 𝜉 ‖ . (4)
‖ 0+ ‖𝐿𝑝
5
Theorem 2. Let 𝜓 (⋅) be an increasing and positive monotone function on (𝑎, 𝑏), having a continuous derivative 𝜓 (⋅) ≠ 0,

on (𝑎, 𝑏). If 0 < 𝛼 ≤ 1 and 0 ≤ 𝛽 ≤ 1, then


𝑏
( )
𝑏 ( )
𝐇 𝛼,𝛽;𝜓 ′ 𝐇 𝛼,𝛽;𝜓 𝜃 (𝑥)
D𝑎+ 𝜉 (𝑥) 𝜃 (𝑥) dx = 𝜉 (𝑥) 𝜓 (𝑥) D𝑎+ dx
∫ ∫ 𝜓 ′ (𝑥)
𝑎 𝑎

for any 𝜉 ∈ 𝐴𝐶 1 and 𝜃 ∈ 𝐶 1 satisfying the boundary conditions 𝜉 (𝑎) = 𝜉 (𝑏) = 0.


From Eq.(1), we have
𝑇 ( ) 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2 ( )
𝐇
D𝛼,𝛽;𝜓 | D0+ 𝜉 (𝑥)| 𝐇 𝛼,𝛽;𝜓
D0+ 𝜉 (𝑥) dx = 𝜆 |𝜉 (𝑥)|𝑝−2 𝜉 (𝑥) + 𝐛 (𝑥) |𝜉 (𝑥)|𝑞−1 𝜉 (𝑥) dx.
∫ 𝑇 | | ∫
0 0
For 𝜂 ∈ ∞
𝐶0 ([0, 𝑇 ], ℝ), we have
𝑇 ( ) 𝑇 𝑇
𝐇 𝛼,𝛽;𝜓 |𝐇 𝛼,𝛽;𝜓 |𝑝−2 𝐇 𝛼,𝛽;𝜓
D𝑇 | D0+ 𝜉 (𝑥)| D0+ 𝜉 (𝑥) 𝜂 (𝑥) dx = 𝜆 |𝜉 (𝑥)|𝑝−2 𝜉 (𝑥) 𝜂(𝑥)dx + 𝐛 (𝑥) |𝜉 (𝑥)|𝑞−1 𝜉 (𝑥) 𝜂(𝑥)dx. (5)
∫ | | ∫ ∫
0 0 0
Using the Theorem 2, we get
𝑇 ( ) 𝑇 ( )
|𝐇 𝛼,𝛽;𝜓 |𝑝−2 |𝐇 𝛼,𝛽;𝜓 |𝑝−2 𝜂 (𝑥)
𝐇
D𝛼,𝛽;𝜓 | D0+ 𝜉 (𝑥)| 𝐇 𝛼,𝛽;𝜓
D0+ 𝜉 (𝑥) 𝜂 (𝑥) dx = | D0+ 𝜉 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) 𝜓 ′ (𝑥) 𝐇
D𝛼,𝛽;𝜓 dx.
∫ 𝑇 | | ∫ | | 0+ 0+ 𝜓 ′ (𝑥)
0 0
J. Vanterler da C. Sousa ET AL 3

( )
𝜂 (𝑥) 1 𝐇 𝛼,𝛽;𝜓
If 𝐇
D𝛼,𝛽;𝜓
0+
= D0+ 𝜂 (𝑥) , ∀𝑥 ∈ [0, 𝑇 ], then Eq.(5), can be rewritten as
𝜓 ′ (𝑥) 𝜓 ′ (𝑥)
𝑇 𝑇 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
| D0+ 𝜉 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) dx =𝜆 |𝜉 (𝑥)|𝑝−2
𝜉 (𝑥) 𝜂(𝑥)dx + 𝐛 (𝑥) |𝜉 (𝑥)|𝑞−1 𝜉 (𝑥) 𝜂(𝑥)dx.
∫ | | 0+ 0+ ∫ ∫
0 0 0
Consider 𝜂 = 𝜉, yields
𝑇 𝑇 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉 (𝑥)| dx = 𝜆 |𝜉 (𝑥)|𝑝 dx + 𝐛 (𝑥) |𝜉 (𝑥)|𝑞+1 dx. (6)
∫ | | ∫ ∫
0 0 0
So, from Eq.(6), we have the functional associated to problem Eq.(1), 𝐄𝜆 ∶ ℍ𝛼,𝛽;𝜓
𝑝 → ℝ, given by 5
𝑇 𝑇 𝑇
1 |𝐇 𝛼,𝛽;𝜓 |𝑝 𝜆 1
𝐄𝜆 (𝜉) = | D 𝜉 (𝑥)| dx − |𝜉 (𝑥)|𝑝 dx − 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx. (7)
𝑝 ∫ | 0+ | 𝑝∫ 𝑞+1∫
0 0 0

We say that 𝜉, 𝜂 ∈ ℍ𝛼,𝛽;𝜓


𝑝 is a solution of the problem (1), if
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
⟨𝐄′𝜆 (𝜉) , 𝜂⟩ℍ𝛼,𝛽;𝜓 = | D0+ 𝜉 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) dx
𝑝 ∫ | | 0+ 0+
0
𝑇 𝑇

−𝜆 |𝜉 (𝑥)|𝑝−2
𝜉 (𝑥) 𝜂 (𝑥) dx − 𝐛 (𝑥) |𝜉(𝑥)|𝑞−1 𝜉(𝑥)𝜂(𝑥)dx. (8)
∫ ∫
0 0

Consider the eigenvalue problem associated with 𝜓-fractional 𝑝-Laplacian operator is given by 5
⎧ ( )
𝛼,𝛽;𝜓 |𝐇 𝛼,𝛽;𝜓 |𝑝−2 𝐇 𝛼,𝛽;𝜓
⎪ D𝑇
𝐇
| D0+ 𝜉(𝑥)| D0+ 𝜉(𝑥) = 𝜆|𝜉(𝑥)|𝑝−2 𝜉(𝑥), for 𝑥 ∈ [0, 𝑇 ]
⎨ | | (9)
⎪ 𝐈𝛽(𝛽−1);𝜓 𝜉(0) = 𝐈𝛽(𝛽−1);𝜓 𝜉(𝑇 ) = 0.
⎩ 0+ 𝑇

Note that,
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉(𝑥)| dx
∫ | |
0
𝜆1 = min , 𝜉 ≠ 0.
𝛼,𝛽;𝜓
𝜉∈ℍ𝑝 𝑇

|𝜉(𝑥)|𝑝 dx

0
𝑇

𝐛 (𝑥) ||𝜙1 || dx, where 𝜙1 is the positive principal eigenvalue


𝑞+1
The direction of bifurcation being determined by the sing

0
corresponding to 𝜆1 .
𝑇

𝐛 (𝑥) ||𝜙1 || dx with


𝑞+1
The main purpose of this paper to discuss bifurcation from infinity using the sign of the integral

0
1 < 𝑞 < 𝑝 − 1 and notice how these facts are related to the properties of the Nehari manifold for the problem (1). This paper is
a continuation of the paper 5 . In the rest, the paper is divided as follows. In section 2, we present the definition of the 𝜓-Hilfer
fractional derivative and some properties of the Nehari manifold and fibering maps. In section 3, we discuss the bifurcation from
infinity.

2 PRELIMINARIES AND NEHARI MANIFOLD

Let 𝑛 − 1 < 𝛼 < 𝑛, with 𝑛 ∈ ℕ, 𝐼 = [𝑎, 𝑏] be an interval such that −∞ ≤ 𝑎 < 𝑏 ≤ ∞ and let 𝜉, 𝜓 ∈ 𝐶 𝑛 ([𝑎, 𝑏], ℝ) be two
functions, such that 𝜓 is increasing and 𝜓 ′ (𝑥) ≠ 0, for all 𝑥 ∈ 𝐼. The left-sided 𝜓-Hilfer fractional derivative 𝐇 𝔻𝛼,𝛽;𝜓
𝑎+ (⋅) of
4 J. Vanterler da C. Sousa ET AL

function 𝜉, of order 𝛼 and type 0 ≤ 𝛽 ≤ 1 is defined by 18,19


( )𝑛
𝛼,𝛽;𝜓 𝛽(𝑛−𝛼);𝜓 1 𝑑 (1−𝛽)(𝑛−𝛼);𝜓
𝐇
D0+ 𝜉(𝑥) ∶= 𝐈0+ 𝐈0+ 𝜉(𝑥), (10)
𝜓 ′ (𝑥) 𝑑𝑡
where 𝐈𝜃;𝜓
0+
(⋅), with 𝜃 = 𝛽(𝑛 − 𝛼) or 𝜃 = (1 − 𝛽)(𝑛 − 𝛼), are the 𝜓-Riemann-Liouville fractional integrals. The right-sided
𝜓-Hilfer fractional derivatives 𝐇 D𝛼,𝛽;𝜓
𝑇
(⋅) is defined in an analogous form 18,19 .
Theorem 3. 20 Suppose that 𝑀 is a closed symmetric 𝐶 1 -submanifold of a real Banach space 𝑋 and 0 ∉ 𝑀. Suppose also that
𝑓 ∈ 𝐶 1 (𝑀, ℝ) is even and bounded below. Define
𝑐𝑗 = inf sup 𝑓 (𝑥)
Λ∈Γ𝑗 𝑥∈Λ
{ ∑ }
where Γ𝑗 = 𝐴 ⊂ 𝑀 ∶ 𝐴 ∈ , 𝛾(𝐴) ≥ 𝑗 and 𝐴 is compact . If Γ𝑘 ≠ ∅ for some 𝑘 ≥ 1 and if 𝑓 satisfies (𝑃 𝑆)𝑐 (Palais–Smale)
for all 𝑐 = 𝑐𝑗 , 𝑗 = 1, ⋯ , 𝑘, then 𝑓 has at least 𝑘 distinct pairs of critical points.
Lemma 1. 5 Let 0 < 𝛼 < 1 and 0 ≤ 𝛽 ≤ 1.
1. The functional 𝐄𝜆 is well defined(on ℍ𝑝𝛼,𝛽;𝜓 )
.
2. The functional 𝐄𝜆 is of class 𝐶 1 ℍ𝛼,𝛽;𝜓
𝑝 ,ℝ .
( )
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
Now we analyze the properties of the 𝜓-Hilfer fractional 𝑝-Laplace operator 𝐇
D𝛼,𝛽;𝜓 𝛼,𝛽;𝜓
| D0+ 𝜉(𝑥)| 𝐇 D0+ 𝜉(𝑥) .
𝑇 | |
Consider the following functional
𝑇
1 |𝐇 𝛼,𝛽;𝜓 |𝑝
(𝜉) = | D0+ 𝜉(𝑥)| dx.
𝑝∫ | |
0
( ) ( )
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
Note that  ∈ 𝐶 1 ℍ𝛼,𝛽;𝜓 , ℝ and the 𝜓-Hilfer fractional 𝑝-Laplace operator 𝐇 D𝛼,𝛽;𝜓 𝛼,𝛽;𝜓
| D0+ 𝜉(𝑥)| 𝐇 D0+ 𝜉(𝑥) , is the
𝑝 𝑇 | |
derivative of  in the weak sense, namely
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
⟨ (𝜉) , 𝜂⟩ℍ𝑝𝛼,𝛽;𝜓 =

| D0+ 𝜉 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) dx
∫ | | 0+ 0+
0

for any 𝜉, 𝜂 ∈ ℍ𝛼,𝛽;𝜓


𝑝 .
( )∗
Theorem 4. 1.  ′ ∶ ℍ𝛼,𝛽;𝜓
𝑝 → ℍ𝛼,𝛽;𝜓
𝑝 is bounded and strictly monotone operator.
⟨ ( ) ⟩
2.  is a mapping of type (𝐒+ ), namely , if 𝜉𝑛 ⇀ 𝜉 in ℍ𝛼,𝛽;𝜓

𝑝 and lim sup  ′ 𝜉𝑛 , 𝜉𝑛 − 𝜉 ≤ 0, then 𝜉𝑛 → 𝜉 in ℍ𝛼,𝛽;𝜓
𝑝 .
𝑛→∞

Proof. Indeed, for a 𝜉 ∈ ℍ𝑝𝛼,𝛽;𝜓 , we obtain


‖ ′ (𝜉)‖( 𝛼,𝛽;𝜓 )∗ = |⟨ ′ ⟩|
‖ ‖ ℍ𝑝 sup |  (𝜉) , 𝜂 |
‖𝜂‖ 𝛼,𝛽;𝜓 ≤1
| |
ℍ𝑝

1∕𝑞 1∕𝑝
⎛ 𝑇 ⎞ ⎛ 𝑇 ⎞
| |(𝑝−1)𝑞 ⎟ ⎜ ||𝐇 D𝛼,𝛽;𝜓 𝜉 (𝑥)|| dx⎟
𝑝
≤ sup ⎜ |𝐇 D𝛼,𝛽;𝜓 𝜉 (𝑥) | dx
‖𝜂‖ 𝛼,𝛽;𝜓 ⎜∫ | | ⎟ ⎜∫ | 0+ | ⎟
0+
ℍ𝑝 ⎝0 ⎠ ⎝0 ⎠
𝑝∕𝑞
≤ ‖𝜉‖ .
ℍ𝛼,𝛽;𝜓
𝑝

Furthermore, for any 𝜉, 𝜂 ∈ ℍ𝑝𝛼,𝛽;𝜓 by direct computation, yields


𝑇
⟨ ⟩ 𝑝 𝑝 |𝐇 𝛼,𝛽;𝜓 |𝑝−2
′ ′
𝐼 (𝜉) − 𝐼 (𝜂) , 𝜉 − 𝜂 = ‖𝜉‖ + ‖𝜂‖ − | D 𝜉 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) dx
ℍ𝛼,𝛽;𝜓
𝑝 ℍ𝛼,𝛽;𝜓
𝑝 ∫ | 0+ | 0+ 0+
0
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
− | D0+ 𝜂 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) dx.
∫ | | 0+ 0+
0
J. Vanterler da C. Sousa ET AL 5

By Holder inequality, it holds that


𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
| D0+ 𝜉 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) dx ≤ ‖𝜉‖𝑝−1
𝛼,𝛽;𝜓 ‖𝜂‖ℍ𝑝
𝛼,𝛽;𝜓
∫ | | 0+ 0+ ℍ𝑝
0

and
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
| D0+ 𝜂 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) dx ≤ ‖𝜂‖𝑝−1
𝛼,𝛽;𝜓 ‖𝜉‖ℍ𝑝
𝛼,𝛽;𝜓 .
∫ | | 0+ 0+ ℍ𝑝
0
Therefore, we obtain
( )( )
⟨ ⟩
 ′ (𝜉) −  ′ (𝜂) , 𝜉 − 𝜂 ≥ ‖𝜉‖𝑝−1 𝑝−1
𝛼,𝛽;𝜓 − ‖𝜂‖ 𝛼,𝛽;𝜓 ‖𝜉‖ℍ𝛼,𝛽;𝜓 − ‖𝜂‖ℍ𝛼,𝛽;𝜓 .
ℍ𝑝 ℍ𝑝 𝑝 𝑝

{ } { ( )} ( )
Let 𝜉𝑛 𝑛∈ℕ ⊂ ℍ𝛼,𝛽;𝜓
𝑝 such that  𝜉𝑛 is bounded and  ′ 𝜉𝑛 → 0 as 𝑛 → ∞. Then there exists a constant 𝐿 > 0 such that
| ( )| ‖ ( )‖
| 𝜉𝑛 | ≤ 𝐿 and ‖ 𝜉𝑛 ‖( 𝛼,𝛽;𝜓 )∗ ≤ 𝐿. (11)
| | ‖ ‖ ℍ𝑝
Therefore, if 𝜉 ≠ 𝜂 then inequality (11) implies that  ′ is strictly monotone.
{ }
2. Let 𝜉𝑛 𝑛∈ℕ be a sequence in ℍ𝛼,𝛽;𝜓
𝑝 such that 𝜉𝑛 → 𝜉 and
⟨ ⟩
lim sup  ′ (𝜉) , 𝜉𝑛 − 𝜉 ≤ 0.
𝑛→+∞

Since ℍ𝛼,𝛽;𝜓
𝑝 is uniformly convex space, weak convergence and norm convergence ( imply )∗strong convergence. Therefore, we
‖ ‖
only need to show that ‖𝜉𝑛 ‖ℍ𝛼,𝛽;𝜓 𝛼,𝛽;𝜓
→ ‖𝜉‖ℍ𝑝 . In fact, since  (𝜉) is an element of ℍ𝑝
𝛼,𝛽;𝜓

, then by weak convergence
𝑝
⟨ ′ ⟩
lim  (𝜉) , 𝜉𝑛 − 𝜉 = 0.
𝑛→+∞

Therefore, using the inequality (11), we have


⟨ ( ) ⟩
0 ≥ lim sup  ′ 𝜉𝑛 − 𝐼 ′ (𝜉) , 𝜉𝑛 − 𝜉
𝑛→+∞
( )( )
≥ lim sup ‖ ‖ ‖𝜉𝑛 ‖ 𝛼,𝛽;𝜓 − ‖𝜉‖ 𝛼,𝛽;𝜓 ≥ 0.
𝑝−1 𝑝−1
𝜉
‖ 𝑛 ‖ 𝛼,𝛽;𝜓 − ‖𝜉‖ 𝛼,𝛽;𝜓 ‖ ‖ℍ𝑝 𝐻𝑝
𝑛→+∞ ℍ𝑝 ℍ𝑝

Then, ‖ ‖
‖𝜉𝑛 ‖ℍ𝛼,𝛽;𝜓
𝑝
→ ‖𝜉‖ℍ𝑝𝛼,𝛽;𝜓 , which implies that 𝜉𝑛 → 𝜉 in ℍ𝛼,𝛽;𝜓
𝑝 . Thus, we concluded the proof.

We start our analysis by considering the eigenvalue problem Eq.(9). We means by a weak solution of Eq.(9), any 𝜉 ∈ ℍ𝛼,𝛽;𝜓
𝑝
such that
𝑇 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2 𝛼,𝛽;𝜓
| D0+ 𝜉 (𝑥)| 𝐇
D0+ 𝜉 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) dx =𝜆 |𝜉 (𝑥)|𝑝−2 𝜉 (𝑥) 𝜂(𝑥)dx (12)
∫ | | 0+ ∫
0 0

for every 𝜂 ∈ ℍ𝛼,𝛽;𝜓


𝑝 . It is easy to see that, if (𝜉, 𝜆) is a solution of Eq.(9) and 𝜉 ≠ 0, then
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝
| D 𝜉 (𝑥)| dx
∫ | 0+ |
0
𝜆 = 𝜆 (𝜉) = , (13)
𝑇

|𝜉 (𝑥)|𝑝 dx

0

and 𝜆 > 0. First, we define , Ψ ∶ ℍ𝛼,𝛽;𝜓


𝑝 → ℝ by
𝑇
1 |𝐇 𝛼,𝛽;𝜓 |𝑝
 (𝜉) ∶= | D0+ 𝜉 (𝑥)| dx
𝑝∫ | |
0
6 J. Vanterler da C. Sousa ET AL

and
𝑇
1
Ψ (𝜉) ∶= |𝜉 (𝑥)|𝑝 dx,
𝑝∫
0
( )
so using Lemma 1 we have , Ψ ∈ 𝐶 1 ℍ𝑝𝛼,𝛽;𝜓 , ℝ ,  and Ψ are even and
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝−2
 ′ (𝜉) 𝜂 ∶= | D 𝜉 (𝑥)| 𝐇
D𝛼,𝛽;𝜓 𝜉 (𝑥) 𝐇
D𝛼,𝛽;𝜓 𝜂 (𝑥) dx,
∫ | 0+ | 0+ 0+
0

𝑇

Ψ (𝜉) 𝜂 ∶= |𝜉 (𝑥)|𝑝−2 𝜉 (𝑥) 𝜂 (𝑥) dx

0

for all 𝜉, 𝜂 ∈ ℍ𝛼,𝛽;𝜓


𝑝 . Also, for any 𝑥 > 0 set
{ }
𝑆 (𝑥) ∶=  −1 (𝑥) = 𝜉 ∈ ℍ𝛼,𝛽;𝜓
𝑝
∶  (𝜉) = 𝑥

then 𝑆 (𝑥) is a 𝐶 1 sub-manifold of ℍ𝑝𝛼,𝛽;𝜓 since 𝑥 is a regular value of . Its obvious that, the functional Ψ𝑥 ∶= Ψ𝑆(𝑥) ∶ 𝑆 (𝑥) → ℝ
defined on 𝑆 (𝑥) is a 𝐶 1 functional. Therefore, if 𝜉 is a critical point of Ψ𝑥 on 𝑆 (𝑥), then (𝜉, 𝜆) is a solution of Eq.(9) (see 21 ,
( )
pp.292). Also, we know that  ′ is of type 𝐒+ , i.e., if 𝜉𝑛 → 𝜉 in ℍ𝛼,𝛽;𝜓𝑝 and
⟨ ′( ) ⟩
lim sup  𝜉𝑛 −  ′ (𝜉) , 𝜉𝑛 − 𝜉 ≤ 0,
𝑛→∞

then 𝜉𝑛 → 𝜉 in ℍ𝛼,𝛽;𝜓
𝑝 (see
Theorem 4). Furthermore Ψ′ is completely continuous and by a standard argument, it follows that Ψ𝑥
satisfies Palais–Smale condition on 𝑆 (𝑥).
Now, we define the following
∑ { }
= 𝐴 ⊂ ℍ𝛼,𝛽;𝜓
𝑝
∖ {0} ∶ 𝐴 is compact and − 𝐴 = 𝐴
∑ { ∑ }
= 𝐴∈ ∶ 𝛾 (𝐴) ≥ 𝑛 , 𝑛 = 1, 2, ..., where 𝛾 (𝐴) is the genus of 𝐴
𝑛
and
𝐶𝑛 (𝑥) = sup

inf Ψ (𝜉) , 𝑛 = 1, 2, ...
𝐴∈ 𝜉∈𝐴
𝑛 ,𝐴⊂𝑆(𝑥)

clearly, 𝐶𝑛 (𝑥) > 0 and 𝐶1 (𝑥) ≥ 𝐶2 (𝑥) ≥ ⋅ ⋅ ⋅ ≥ 𝐶𝑛 (𝑥) ≥ ⋅ ⋅ ⋅.


By the Ljusternik-Schrirelman theory on 𝐶 1 −manifolds 20 , we have the following theorem.
Theorem 5. 1. 𝐶𝑛 (𝑥) → 0 as 𝑛 → ∞.
2. 𝐶𝑛 (𝑥) is a critical value of Ψ𝑥 on 𝑆 (𝑥) and the Ljusternik-Schrirelman result holds.
Set
Δ = {𝜆 ∈ ℝ ∶ 𝜆 is an eigenvalue of Eq.(9)} ,
{ }
Θ𝑛 (𝑥) = 𝜉 ∈ 𝑆 (𝑥) ∶ 𝜉 is a critical point of Ψ𝑥 and Ψ𝑥 (𝜉) = 𝐶𝑛 (𝑥) ,
{ }
Δ𝑛 (𝑥) = 𝜆 (𝜉) ∶ 𝜉 ∈ Θ𝑛 , where 𝜆 (𝜉) is defined in Eq.(13).
Hence, for each 𝑛, Θ𝑛 (𝑥) and Δ𝑛 (𝑥) are nonempty, and Δ𝑛 (𝑥) ⊂ Δ.
Let 𝜉𝑛 ∈ Θ𝑛 (𝑥), then
𝑇
( ) |𝐇 𝛼,𝛽;𝜓 |𝑝
 𝜉𝑛 ∶= | D0+ 𝜉𝑛 (𝑥)| dx = 𝑥
∫ | |
0
and
𝑇
( )
Ψ 𝜉𝑛 ∶= |𝜉 (𝑥)|𝑝 dx = 𝐶𝑛 (𝑥) ,
∫ | 𝑛 |
0
J. Vanterler da C. Sousa ET AL 7

and this implies that


𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉𝑛 (𝑥)| dx
∫ | |
( ) 0 𝑥
𝜆𝑛 = 𝜆 𝜉𝑛 = = → +∞
𝑇 𝐶𝑛 (𝑥)
|𝜉 (𝑥)|𝑝 dx
∫ | 𝑛 |
0
( )
as 𝑛 → ∞ and 𝜉𝑛 , 𝜆𝑛 is a solution of Eq.(9), for any 𝑛 ∈ ℕ. Therefore, we can get the following result.
Theorem 6. Δ is nonempty infinite set and sup Δ = +∞.
Then the problem Eq.(9) has a unique of eigenvalue where 𝜆1 is the first eigenvalue of Eq.(9) and
𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉 (𝑥)| dx
∫ | |
0
𝜆1 = min , 𝜉 ≠ 0.
𝛼,𝛽;𝜓
𝜉∈ℍ𝑝 𝑇

|𝜉 (𝑥)|𝑝 dx

0

Now, we present some results about the Nehari manifold structure and fibering map analysis, related to problem (1). Let us
consider real functions of positive variables as follows: 𝚵𝜉 (𝑡) ∶ 𝑡 → 𝐄𝜆 (𝑡𝜉) (𝑡 > 0), and these functions are known as fibering
𝛼,𝛽;𝜓 𝛼,𝛽;𝜓
maps for each 𝜉 ∈ ℍ𝛼,𝛽;𝜓
𝑝 and from Eq.(7), yields (note that 𝐇 D0+ (𝑎𝜉) = 𝑎 𝐇 D0+ (𝜉)) 5
𝑇 𝑇 𝑇
1 |𝐇 𝛼,𝛽;𝜓 |𝑝 𝜆 1
𝚵𝜉 (𝑡) = | D 𝑡𝜉 (𝑥)| dx − |𝑡𝜉 (𝑥) |𝑝 dx − 𝐛 (𝑥) |𝑡𝜉 (𝑥) |𝑞+1 dx
𝑝 ∫ | 0+ | 𝑝∫ 𝑞+1∫
0 0 0
𝑇 𝑇
( )
𝑡𝑝 |𝐇 𝛼,𝛽;𝜓 |𝑝 𝑡𝑞+1
= | D0+ 𝜉 (𝑥)| − 𝜆|𝜉 (𝑥) |𝑝 dx − 𝐛 (𝑥) |𝜉 (𝑥) |𝑞+1 dx. (14)
𝑝∫ | | 𝑞+1∫
0 0

Applying the derivative on both sides of Eq.(14) (in relation to 𝑡), we have
𝑇 𝑇
( )
|𝐇 𝛼,𝛽;𝜓 |𝑝
𝚵′𝜉 (𝑡) =𝑡 𝑝−1
| D0+ 𝜉(𝑥)| − 𝜆|𝜉(𝑥)|𝑝 dx − 𝑡𝑞 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx. (15)
∫ | | ∫
0 0

Finally, applying the derivative on both sides of Eq.(15) (in relation to 𝑡), we obtain
𝑇 𝑇
( )
|𝐇 𝛼,𝛽;𝜓 |𝑝
𝚵′′𝜉 (𝑡) = (𝑝 − 1)𝑡 𝑝−2
| D0+ 𝜉(𝑥)| − 𝜆|𝜉(𝑥)|𝑝 dx − 𝑞𝑡𝑞−1 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx. (16)
∫ | | ∫
0 0

It is natural to divide the Nehari manifold into three distinct sets. Note that Eq.(15) and Eq.(16), yields
𝑇

𝚵′′𝜉 (𝑡) =𝑡 𝑞−1


[(𝑝 − 1) − 𝑞] 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx

0

and as 𝑡 > 0, 1 < 𝑞 < 𝑝 − 1, we have


𝑇 𝑇

𝚵′′𝜉 (𝑡) > 0 ⇔ 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx > 0, 𝚵′′𝜉 (𝑡) < 0 ⇔ 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx < 0
∫ ∫
0 0

and
𝑇

𝚵′′𝜉 (𝑡) =0⇔ 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx = 0.



0
8 J. Vanterler da C. Sousa ET AL

We can divide 𝐍𝜆 into three sets:


⎧ 𝑇 ⎫ ⎧ 𝑇 ⎫
⎪ ⎪ ⎪ ⎪
𝐍+𝜆 = ⎨𝜉 ∈ 𝐍𝜆 ; 𝐛 (𝑥) |𝜉(𝑥)| dx > 0⎬ , 𝐍𝜆 = ⎨𝜉 ∈ 𝐍𝜆 ;
𝑞+1 −
𝐛 (𝑥) |𝜉(𝑥)| dx < 0⎬
𝑞+1

⎪ ∫ ⎪ ⎪ ∫ ⎪
⎩ 0 ⎭ ⎩ 0 ⎭
and
⎧ 𝑇 ⎫
⎪ ⎪
𝐍0𝜆 = ⎨𝜉 ∈ 𝐍𝜆 ; 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx = 0⎬ ,
⎪ ∫ ⎪
⎩ 0 ⎭
where 𝐍𝜆 , 𝐍𝜆 and 𝐍𝜆 correspond to the points of local minimum, local maximum and inflection of 𝚵𝜉 , respectively.
+ − 0

The analysis of the behavior of 𝚵𝜉 is in relation to the sign of


𝑇 𝑇
( )
|𝐇 𝛼,𝛽;𝜓 |𝑝
𝚯(𝜉) = | D0+ 𝜉(𝑥)| − 𝜆|𝜉(𝑥)|𝑝 dx and 𝚼(𝜉) = 𝐛 (𝑥) |𝜉(𝑥)|𝑞+1 dx.
∫ | | ∫
0 0
5
Let 𝜉 ∈ ℍ𝛼,𝛽;𝜓
𝑝 and 𝚵′𝜉 (𝑡) = 0, and we conclude that :

1. If 𝚯(𝜉) and 𝚼(𝜉) have the same sign, then 𝚵𝜉 has a unique critical point,
1
⎛ 𝑇 ⎞ 𝑝−(1+𝑞)
⎜ 𝐛 (𝑥) |𝜉(𝑥)| dx
𝑞+1 ⎟
⎜ ∫ ⎟
⎜ 0 ⎟
𝑡𝜉 = ⎜ ⎟ .
𝑇
⎜ ( ) ⎟
⎜ |𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉(𝑥)| − 𝜆|𝜉(𝑥)|𝑝 dx ⎟
⎜∫ | | ⎟
⎝0 ⎠
This critical point is a minimum when 𝑡𝜉 𝜉 ∈ 𝐍+𝜆 if, and only if, 𝚼(𝜉) > 0. It is a maximum point when 𝑡𝜉 𝜉 ∈ 𝐍−𝜆 if, and
only if, 𝚼(𝜉) < 0.

2. If 𝚯(𝜉) and 𝚼(𝜉) have opposite signs, then 𝚵𝜉 has no inflection points, so there are no multiples of 𝜉 in 𝐍𝜆 .

Therefore, we define (here ‖ . ‖ = ‖ . ‖ℍ𝛼,𝛽;𝜓


𝑝
)

⎧ 𝑇 ⎫
⎪ ( ) ⎪
|𝐇 𝛼,𝛽;𝜓 |𝑝
𝐋+ (𝜆) = ⎨𝜉 ∈ ℍ𝑝 ; ‖𝜉‖ = 1 and
𝛼,𝛽;𝜓
| D0+ 𝜉(𝑥)| − 𝜆|𝜉(𝑥)| dx > 0⎬ and
𝑝
∫ | |
⎪ ⎪
⎩ 0 ⎭
⎧ 𝑇 ⎫
⎪ ⎪
𝚼+ = ⎨𝜉 ∈ ℍ𝑝 ; ‖𝜉‖ = 1 and
𝛼,𝛽;𝜓
𝐛 (𝑥) |𝜉(𝑥)| dx > 0⎬ .
𝑞+1

⎪ ∫ ⎪
⎩ 0 ⎭
Similarly, we can define 𝐋− (𝜆) and 𝚼− when we replace “> 0” by “< 0” and 𝐋0 (𝜆) and 𝚼0 when we replace “> 0” by “= 0”.
We have the following results:

1. If 𝜉 ∈ 𝐋+ (𝜆) ∩ 𝚼+ , then 𝑡 → 𝚵𝜉 (𝑡) has a local minimum at 𝑡 = 𝑡𝜉 and 𝑡𝜉 𝜉 ∈ 𝐍+𝜆 .

2. If 𝜉 ∈ 𝐋− (𝜆) ∩ 𝚼− , then 𝑡 → 𝚵𝜉 (𝑡) has a local maximum at 𝑡 = 𝑡𝜉 and 𝑡𝜉 𝜉 ∈ 𝐍−𝜆 .

3. If 𝜉 ∈ 𝐋+ (𝜆) ∩ 𝚼− , then 𝑡 → 𝚵𝜉 (𝑡) is strictly increasing and no multiple of 𝜉 is in 𝐍𝜆 .

4. If 𝜉 ∈ 𝐋− (𝜆) ∩ 𝚼+ , then 𝑡 → 𝚵𝜉 (𝑡) is strictly decreasing and no multiple of 𝜉 is in 𝐍𝜆 .


Theorem 7. 5 Suppose that there exists 𝜆, ̄ such that for any 𝜆 < 𝜆,̄ we have 𝐋− (𝜆) ⊆ 𝚼− . Then, for all 𝜆 < 𝜆,
̄ we get

1. 𝐋0 (𝜆) ⊆ 𝚼− and, so, 𝐋0 (𝜆) ∩ 𝚼0 = ∅,

2. 𝐍+𝜆 is bounded,
J. Vanterler da C. Sousa ET AL 9

3. 0 ∉ 𝐍−𝜆 and 𝐍−𝜆 is closed,

4. 𝐍+𝜆 ∩ 𝐍−𝜆 = ∅.

3 BIFURCATION FROM INFINITY

Since 𝐋− (𝜆) = ∅ for 𝜆 < 𝜆1 , from Theorem 1, that there is a minimizer of 𝐄𝜆 in 𝐍+𝜆 whenever 𝜆 < 𝜆1 . Now, we study the
asymptotic behavior of the minimizes on 𝐍+𝜆 as 𝜆 → 𝜆−1 .
𝑇

𝐛 (𝑥) ||𝜙1 || dx > 0. Then lim− inf+ 𝐄𝜆 (𝜉) = −∞.


𝑞+1
Theorem 8. Suppose that
∫ 𝜆→𝜆1 𝜉∈𝐍𝜆
0

𝑇 𝑇 𝑇
( ) ( )
|𝐻 𝛼,𝛽;𝜓 |𝑝
𝐛 (𝑥) ||𝜙1 |
|𝑞+1 | D0+ 𝜙1 | − 𝜆 ||𝜙1 || dx = 𝜆1 − 𝜆 |𝜙 |𝑝 dx we have that 𝜙1 ∈ 𝐋+ (𝜆) ∩
𝑝
Proof. Since dx > 0 and
∫ ∫ | | ∫ | 1|
0 0 0
Υ+ for all 𝜆1 > 𝜆 and so 𝑡𝜙1 𝜙1 ∈ 𝐍+𝜆 .
Now
( ) 𝑇
( )
( ) 1 1 ( )𝑝 |𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜙1 | − 𝜆 ||𝜙1 || dx
𝑝
𝐄𝜆 𝑡𝜙1 𝜙1 = − 𝑡𝜙
𝑝 𝑞+1 ∫ | |
0
𝑝
⎛ 𝑇 ⎞ 𝑝−(𝑞+1)
⎜ 𝑏 (𝑥) ||𝜙1 || dx
𝑞+1 ⎟
( )⎜ ∫ ⎟ 𝑇
( )
1 1 ⎜ ⎟ |𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜙1 | − 𝜆 ||𝜙1 || dx
0 𝑝
𝑝 𝑞 + 1 ⎜⎜ 𝑇 ⎟
= −
∫ | |
( ) ⎟
|𝐇 𝛼,𝛽;𝜓 | 𝑝 0
⎜ | D0+ 𝜙1 | − 𝜆 ||𝜙1 || dx ⎟
𝑝
⎜∫ | | ⎟
⎝0 ⎠
𝑝 𝑞+1
( )⎛ 𝑇 ⎞ 𝑝−(𝑞+1) ⎛ 𝑇 ⎞ (𝑞+1)−𝑝 (
1 1 ) 𝑞+1
= − ⎜ 𝑏 (𝑥) |𝜙1 | dx⎟
| | 𝑞+1 ⎜ 𝑏 (𝑥) |𝜙1 |𝑞+1 dx⎟ 𝜆 − 𝜆 (𝑞+1)−𝑝
.
𝑝 𝑞 + 1 ⎜∫ ⎟ ⎜∫ | | ⎟ 1
⎝0 ⎠ ⎝0 ⎠
(17)
( ) ( )
Taking limit 𝜆 → 𝜆−1 in Eq.(17), we have (note 𝑞 < 𝑝−1) lim 𝐄𝜆 𝑡𝜙1 𝜙1 → −∞. So, we have inf𝜆 𝐄𝜆 (𝜉) ≤ 𝐄𝜆 𝑡𝜙1 𝜙1 → −∞
𝜆→𝜆1 𝜉∈𝐍+
when 𝜆 → 𝜆−1 .

𝐛 (𝑥) ||𝜙1 || dx > 0. Then, for each 𝜆 < 𝜆1 there exists a minimizer 𝜉𝜆 in 𝐍+𝜆 such that lim− ‖ 𝜉𝜆 ‖ =
𝑞+1
Corollary 1. Suppose that
∫ 𝜆→𝜆 ‖ ‖ 1
0
∞.

Proof. Let 𝜆 < 𝜆1 and 𝜉𝜆 ∈ 𝐍+𝜆 a minimizer. Then, we have


𝑇 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝 ||𝐇 𝛼,𝛽;𝜓 |𝑝 |
lim− ‖

‖𝑝
𝜉𝜆 ‖ = lim− | D0+ 𝜉𝜆 | dx ≥ lim− || D0+ 𝜉𝜆 | − 𝜆 |𝜉|𝑝 | dx
𝜆→𝜆1 𝜆→𝜆1 ∫ | | 𝜆→𝜆1 ∫ || | |
0 0
( } )−1 {
1 1
− ≥
lim− inf+ 𝐄𝜆 (𝜉) → ∞
𝑝 𝑞+1 𝜆→𝜆1 𝜉∈𝑆𝜆
( ) { }
1 1
because 𝑞 < 𝑝 − 1 implies that − < ∞ and lim− inf+ 𝐄𝜆 (𝜉) → −∞ (see Theorem 8).
𝑝 𝑞+1 𝜆→𝜆1 𝜉∈𝐍𝜆
10 J. Vanterler da C. Sousa ET AL

𝐛 (𝑥) ||𝜙1 || dx < 0. In this case the hypothesis of Theorem 7 is maintained


𝑞+1
Let us now turn our attention to the case where

0
in some way, the right of 𝜆 = 𝜆1 . More precisely, we have the following result:
𝑇 𝑇

Lemma 2. Suppose that 𝐛 (𝑥) ||𝜙1 || dx < 0. Then, there exists 𝛿1 , 𝛿2 > 0 such that 𝜉 ∈ 𝐋− (𝜆) implies 𝐛 (𝑥) ||𝜙1 || dx ≤
𝑞+1 𝑞+1
∫ ∫
0 0
−𝛿2 whenever 𝜆1 ≤ 𝜆 ≤ 𝜆1 + 𝛿1 .
𝑇
( )
𝐛 (𝑥) ||𝜉𝑛 || dx > 𝛿2𝑛 , whenever 𝜆1 ≤ 𝜆𝑛 ≤
𝑞+1
Proof. Suppose by absurdity that for 𝛿1𝑛 , 𝛿2𝑛 > 0 exists 𝜉𝑛 ∈ 𝐋− 𝜆𝑛 such

0
1 { } 1
𝜆1 + 𝛿1𝑛 . We taking 𝛿1𝑛 = 𝛿2𝑛 = e 𝜆𝑛 𝑛∈ℕ∕{1} a strictly sequence such that 𝜆1 ≤ 𝜆𝑛 ≤ 𝜆1 + . Thus, for each 𝑛 ∈ ℕ∕ {1},
𝑛 𝑛
𝑇
( ) 1
𝐛 (𝑥) ||𝜉𝑛 || dx > − .
𝑞+1
there exists 𝜉𝑛 ∈ 𝐋− 𝜆𝑛 such that
∫ 𝑛
0
Observe that
𝑇 𝑇 𝑇 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝 |𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉 | dx − 𝜆𝑛 |𝜉|𝑝 dx < | D 𝜉 | dx − 𝜆𝑛+1 |𝜉|𝑝 dx
∫ | | ∫ ∫ | 0+ | ∫
0 0 0 0
𝑇 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝
< ⋅⋅⋅< | D 𝜉 | dx − 𝜆1 |𝜉|𝑝 dx.
∫ | 0+ | ∫
0 0
(18)
( ) ( ) ( ) ( ) ( )
Thus, if 𝜉 ∈ 𝐋− 𝜆𝑛+1 implies 𝜉 ∈ 𝐋− 𝜆𝑛 , soon we have 𝐋− 𝜆𝑛 ⊃ 𝐋− 𝜆𝑛+1 ⊃ 𝐋− 𝜆𝑛+2 ⊃ ⋅ ⋅ ⋅. Still, for all 𝜆 > 𝜆1 ,
yields
𝑇 𝑇
( ) ( )
|𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜙1 | − 𝜆 ||𝜙1 || dx = 𝜆1 − 𝜆 |𝜙 |𝑝 dx < 0
𝑝
(19)
∫ | | ∫ | 1|
0 0
implies that 𝜙1 ∈ 𝐋− (𝜆).
( ) ( ) ( )
In addition, it has already been seen that 𝜆 = 𝜆1 implies that 𝐋0 𝜆1 = 𝜙1 , thus 𝐋0 𝜆1 ⊂ 𝐋− (𝜆), ∀𝜆 > 𝜆1 , so 𝐋− 𝜆𝑛 ⊃
( ) ( ) ( ) { } ( ) ( )
𝐋− 𝜆𝑛+1 ⊃ 𝐋− 𝜆𝑛+2 ⊃ ⋅ ⋅ ⋅ ⊃ 𝐋0 𝜆1 . Also note that we can take the sequence 𝜉𝑛 ∈ 𝐋− 𝜆𝑛 such that 𝜉𝑛 ∈ 𝐋− 𝜆𝑛 ,
𝑇 𝑇
( )
∈ 𝐋− 𝜆𝑛+1 , so that 𝜉𝑛 → 𝜙1 . Therefore, ||𝜉𝑛 || → ||𝜙1 || |𝜉 |𝑞+1 dx → |𝜙 |𝑞+1 dx implies that (note 𝑏 ∈ 𝐿∞ )
𝑞+1 𝑞+1
𝜉𝑛+1 ⇒
∫ | 𝑛| ∫ | 1|
0 0
𝑇 𝑇 𝑇
1
𝐛 (𝑥) ||𝜉𝑛 || dx → 𝐛 (𝑥) ||𝜙1 || dx. Thus, we have 𝐛 (𝑥) ||𝜙1 || dx > − , ∀𝑛 ∈ ℕ, which is absurd, since by hypothesis
𝑞+1 𝑞+1 𝑞+1
∫ ∫ ∫ 𝑛
0 0 0
𝑇

𝐛 (𝑥) ||𝜙1 || dx < 0.


𝑞+1
of the theorem,

0
𝑇

𝐛 (𝑥) ||𝜙1 || dx < 0 and 𝛿1 be like Lemma 2. So whenever 𝜆1 ≤ 𝜆 ≤ 𝜆1 + 𝛿1 , there exists


𝑞+1
Theorem 9. Suppose that

0
minimizer 𝜉𝜆 and 𝜂𝜆 of 𝐄𝜆 in 𝐍+𝜆 and 𝐍−𝜆 respectively.

Proof. Note that 𝜙1 ∈ 𝐋− (𝜆) and so we have to 𝐋− (𝜆) ≠ ∅ whenever 𝜆 > 𝜆1 . From Lemma 2, we obtain the hypotheses of
Theorem 1 with 𝜆̂ = 𝜆1 + 𝛿1 and so the result follows.
𝑇

𝐛 (𝑥) ||𝜙1 || dx < 0, the bifurcation of infinity occurs to the right of 𝜆 = 𝜆1 .


𝑞+1
The next result shows that when

0
J. Vanterler da C. Sousa ET AL 11

𝐛 (𝑥) ||𝜙1 || dx < 0. Then we have


𝑞+1
Theorem 10. Suppose

0

1. lim+ inf− 𝐄𝜆 (𝜉) = +∞.


𝜆→𝜆1 𝜉∈𝑆

2. If 𝜆𝑛 → 𝜆+1 and 𝜉𝑛 is a minimizer of 𝐄𝜆𝑛 on 𝐍−𝜆 , then lim ‖ 𝜉𝑛 ‖ = +∞.


𝑛 𝑛→∞ ‖ ‖

Proof. Let be 𝜂 ∈ 𝐍−𝜆 . So 𝑡𝜂 = 𝑡𝜉 (𝜉) for some 𝜉 ∈ 𝐋− (𝜆) ∩ 𝚼− . Now, by means of the Lemma 2, there exists 𝛿1 and 𝛿2 > 0,
𝑇 ( ( ) )
𝜆
such that 𝐛 (𝑥) |𝜉| dx ≤ −𝛿2 whenever 𝜆1 ≤ 𝜆 ≤ 𝜆1 + 𝛿1 and 𝑛𝑜𝑡𝑒 1 −
𝑞+1
<0
∫ 𝜆1
0

𝑇
( ) ( ) 𝑇
|𝐇 𝛼,𝛽;𝜓 |𝑝 𝜆 |𝐇 𝛼,𝛽;𝜓 |𝑝 𝜆 −𝜆
0> | D0+ 𝜉 | − 𝜆 |𝜉|𝑝 dx ≥ 1 − | D 𝜉 | dx = 1
∫ | | 𝜆1 ∫ | 0+ | 𝜆1
0 0

so that
| | 𝑇
) | 𝜆 −𝜆| (
|𝐇 𝛼,𝛽;𝜓 |𝑝 | 1 |
| D0+ 𝜉 | − 𝜆 |𝜉| dx| ≤
𝑝
|.
| | | 𝜆1|∫
| |0
| |
( ( ) )
1 1 1 1
Then, we have note < so − >0
p q+1 q+1 p
( ) 𝑇
( )
1 1 ( )𝑝 |𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉𝑛 (𝑥)| − 𝜆 ||𝜉𝑛 (𝑥)|| dx
𝑝
𝐄𝜆 (𝜂) = − 𝑡𝑛
𝑝 𝑞+1 ∫ | |
0
𝑝

⎛ 𝑇 ⎞ 𝑝−(𝑞+1)
⎜ 𝑏 (𝑥) |𝜉 (𝑥)|𝑞+1 dx⎟
( ) ⎜∫ | 𝑛 | ⎟
1 1 ⎝0 ⎠
= − 𝑞+1
𝑝 𝑞+1
⎛ 𝑇 ( ) ⎞ 𝑝−(𝑞+1)
⎜ |𝐇 𝛼,𝛽;𝜓 |𝑝
| D0+ 𝜉𝑛 (𝑥)| − 𝜆 ||𝜉𝑛 (𝑥)|| dx⎟
𝑝
⎜∫ | | ⎟
⎝0 ⎠
𝑞+1
( )
1 1 ( ) 𝑝 𝜆1𝑝−(𝑞+1)
≤ − 𝛿2 𝑝−(𝑞+1) ( ) 𝑞+1 .
𝑞+1 𝑝
𝜆 − 𝜆1 𝑝−(𝑞+1)
Taking limit 𝜆 → 𝜆+1 , we have inf− 𝐄𝜆 (𝜉) = +∞. Therefore, 𝜂𝜆 is un bounded when 𝜆 → 𝜆1 .
𝜉∈𝐍
2. Is a direct consequence of 1.

CONFLICT OF INTEREST

This work does not have any conflicts of interest

ACKNOWLEDGMENT

J. Vanterler acknowledges the financial support of a PNPD-CAPES (number of process nº88882.305834/2018-01) scholarship
of the Postgraduate Program in Applied Mathematics of IMECC-Unicamp. We are grateful to the anonymous referees for the
suggestions that improved the manuscript.
12 J. Vanterler da C. Sousa ET AL

References

1. Fiscella A., Bisci G. Molica, Servadei R.. Bifurcation and multiplicity results for critical nonlocal fractional Laplacian
problems. Bull. Sci. Math.. 2016;140(1):14–35.

2. Bisci G. Molica, Mugnai D., Servadei R.. On multiple solutions for nonlocal fractional problems via ∇-theorems. Diff.
Integral Equ.. 2017;30(9/10):641–666.

3. Alves Claudianor O., Bisci G. Molica, Ledesma C.. Existence of solutions for a class of fractional elliptic problems on
exterior domains. J. Diff. Equ.. 2020;11:7183–7219.

4. Ledesma C.. Multiplicity result for non-homogeneous fractional Schrodinger-Kirchhoff-type equations in ℝ𝑛 . Adv. Nonlin-
ear Anal.. 2018;7(3):247–257.

5. Sousa J. Vanterler da C., Zuo J., O’Regan Donal. The Nehari manifold for a 𝜓-Hilfer fractional 𝑝-Laplacian,
https://hal.archives-ouvertes.fr/hal-02580245. 2020;.

6. Ledesma C.. Existence and symmetry result for fractional 𝑝-Laplacian in ℝ𝑛 . Commun. Pure Appl. Anal.. 2017;16(1):99.

7. Nyamoradi N., Tersian S.. Existence of solutions for nonlinear fractional order 𝑝-Laplacian differential equations via critical
point theory. Frac. Calc. Appl. Anal.. 2019;22(4):945–967.

8. Bartolo R., Bisci G. Molica. Asymptotically linear fractional 𝑝-Laplacian equations. Annali di Matematica Pura ed
Applicata. 2017;196(2):427–442.

9. Li P.-L., Sun H.-R.. Existence results and bifurcation for nonlocal fractional problems with critical Sobolev exponent.
Comput. Math. Appl.. 2019;78(5):1720–1731.

10. Perera K., Squassina M., Yang Y.. Bifurcation and multiplicity results for critical fractional 𝑝-Laplacian problems. Math.
Nachr.. 2016;289(2-3):332–342.

11. Mishra P. K., Sreenadh K.. Bifurcation and multiplicity of solutions for the fractional Laplacian with critical exponential
nonlinearity. Electron. J. Diff. Equa.. 2016;2016:1–9.

12. Alsulami H., Kirane M., Alhodily S., Saeed T., Nyamoradi N.. Existence and multiplicity of solutions to fractional 𝑝-
Laplacian systems with concave–convex nonlinearities. Bull. Math. Sci.. 2020;10(01):2050007.

13. Ledesma César Torres. Boundary value problem with fractional 𝑝-Laplacian operator. Adv. Nonlinear Anal.. 2016;5(2):133–
146.

14. Ledesma César E. Torres, Nyamoradi N.. Impulsive fractional boundary value problem with 𝑝-Laplace operator. J. Appl.
Math. Comput.. 2017;55(1-2):257–278.

15. Ledesma César Torres. Existence and symmetric result for Liouville–Weyl fractional nonlinear Schrödinger equation.
Commun. Nonlinear Sci. Numer. Simul.. 2015;27(1-3):314–327.

16. Sousa J. Vanterler da C., Pulido M. Aurora P., Oliveira E. Capelas. Existence and regularity of weak solutions for 𝜓-Hilfer
fractional boundary value problem, https://hal.archives-ouvertes.fr/hal-02562931. Mediterr. J. Math. (Accept) 2020;.

17. Sousa J. Vanterler da C., Tavares L., Ledesma César. A variational approach for a problem involving a 𝜓-Hilfer fractiona
operator, https://hal.archives-ouvertes.fr/hal-02615625. J. Appl. Anal. Comput. (Accept) 2020;.

18. Sousa J.Vanterler da C., Oliveira E.. On the 𝜓-Hilfer fractional derivative. Commun. Nonlinear Sci. Numer. Simul..
2018;60:72–91.

19. Sousa J. Vanterler da C., Oliveira E.. Leibniz type rule: 𝜓-Hilfer fractional operator. Commun. Nonlinear Sci. Numer. Simul..
2019;77:305–311.

20. Szulkin A.. Ljusternik-Schnirelmann theory on 𝐶 1 -manifolds. Annales de l’Institut Henri Poincare (C) Non Linear Analysis.
1988;5(2):119–139.
J. Vanterler da C. Sousa ET AL 13

21. Zhang X., Zhu C., Wu Z.. Solvability for a coupled system of fractional differential equations with impulses at resonance.
Boundary Value Probl.. 2013;2013(1):80.

How to cite this article: J. Vanterler da C. Sousa, Nehari manifold and bifurcation for a 𝜓-Hilfer fractional 𝑝-Laplacian, Math.
Meth. Appl. Sci., 2020;1–14.

You might also like