Next Article in Journal
Quasi-Mittag-Leffler Projective Synchronization of Delayed Chaotic Fractional Order Neural Network with Mismatched Parameters
Next Article in Special Issue
Remarks on a New Variable-Coefficient Integro-Differential Equation via Inverse Operators
Previous Article in Journal
Realization of Fractional-Order Current-Mode Multifunction Filter Based on MCFOA for Low-Frequency Applications
Previous Article in Special Issue
The Existence and Stability of Integral Fractional Differential Equations
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Hadamard Fractional Boundary Value Problem on an Infinite Interval at Resonance

1
Department of Computer Science and Engineering, Gh. Asachi Technical University, 700050 Iasi, Romania
2
Department of Mathematics, Gh. Asachi Technical University, 700506 Iasi, Romania
*
Author to whom correspondence should be addressed.
Fractal Fract. 2025, 9(6), 378; https://doi.org/10.3390/fractalfract9060378
Submission received: 14 May 2025 / Revised: 2 June 2025 / Accepted: 9 June 2025 / Published: 13 June 2025

Abstract

This paper addresses the existence of solutions to a Hadamard fractional differential equation of arbitrary order on an infinite interval, subject to integral boundary conditions that incorporate both Riemann–Stieltjes integrals and Hadamard fractional derivatives. Due to the presence of nontrivial solutions in the associated homogeneous boundary value problem, the problem is classified as resonant. The Mawhin continuation theorem is utilized to derive the main findings.

1. Introduction

In this paper, we examine the following fractional differential equation:
D 1 + α H u ( θ ) = f θ , u ( θ ) , D 1 + α n + 1 H u ( θ ) , D 1 + α n + 2 H u ( θ ) , , D 1 + α 1 H u ( θ ) , θ ( 1 , ) ,
subject to the following nonlocal boundary conditions:
u ( j ) ( 1 ) = 0 , j = 0 , , n 3 ; D 1 + α 2 H u ( 1 ) = 1 A D 1 + α 2 H u ( σ ) d P ( σ ) , D 1 + α 1 H u ( ) = 1 B D 1 + α 1 H u ( σ ) d Q ( σ ) ,
where α ( n 1 , n ] , n N , n 3 , f : ( 1 , ) × R n R is a Carathéodory function, D 1 + k H is the Hadamard fractional derivative of order k for k = α , α 1 , , α n + 1 ; A > 1 , B > 1 , and the integrals from the boundary conditions (2) are Riemann–Stieltjes integrals with P : [ 1 , A ] R and Q : [ 1 , B ] R bounded variation functions.
The problem under consideration is resonant in the sense that the associated homogeneous boundary value problem admits nontrivial solutions (refer to assumption ( I 1 ) and Lemma 2 below). By imposing various conditions on the nonlinearity f , we establish the existence of solutions to the boundary value problem (1), (2). The proofs of our main results are based on the coincidence degree theory developed by J. Mawhin, specifically the Mawhin continuation theorem (see Theorem IV.13 from [1,2]). In the subsequent discussion, we briefly review several works that are related to the present study. In [3], the authors studied the Hadamard fractional Equation (1) on the interval of ( 1 , e ) at resonance, supplemented with the following integral boundary conditions:
u ( j ) ( 1 ) = 0 , j = 0 , , n 3 , D 1 + α 1 H u ( 1 ) = 1 e D 1 + α 1 H u ( σ ) d P ( σ ) , D 1 + α 1 H u ( e ) = 1 e D 1 + α 1 H u ( σ ) d Q ( σ ) .
They proved the existence of solutions for the problem (1), (3) on ( 1 , e ) by using the coincidence degree theory. In [4], the authors analyzed a Hadamard fractional differential equation at resonance given by
( D 1 + γ H u ) ( θ ) + f ( θ , u ( θ ) ) = 0 , θ ( 1 , e ) , u ( 1 ) = 0 , u ( e ) = 1 e u ( σ ) d A ( σ ) ,
where γ ( 1 , 2 ) and f : ( 1 , e ) × R R satisfies Carathéodory conditions. The boundary conditions include a Riemann–Stieltjes integral. Using the Mawhin continuation theorem, they established the existence of solutions for problem (4). In [5], the authors investigated the existence of solutions to a Riemann–Liouville fractional boundary value problem at resonance on an infinite interval. Utilizing the coincidence degree theory, they consider the following problem:
D 0 + α u ( θ ) = f ( θ , u ( θ ) , D 0 + α 2 u ( θ ) , D 0 + α 1 u ( θ ) ) , θ ( 0 , ) , u ( 0 ) = 0 , D 0 + α 2 u ( 0 ) = i = 1 m α i D 0 + α 2 u ( ξ i ) , D 0 + α 1 u ( ) = k = 1 n β k D 0 + α 1 u ( η k ) ,
where α ( 2 , 3 ] , 0 < η 1 < η 2 < < η n < , 0 < ξ 1 < ξ 2 < < ξ m < , 0 < η 1 < η 2 < < η n < , and the coefficients are α i , β k R for i = 1 , , m , k = 1 , , n . The function f : ( 0 , ) × R 3 R is assumed to be a Carathéodory function, and D 0 + κ denotes the Riemann–Liouville fractional derivative of order κ for κ { α , α 1 , α 2 } . In [6], the authors examined the multipoint boundary value problem of the following fractional differential equation:
D 0 + α u ( θ ) = f ( θ , u ( θ ) , D 0 + α 1 u ( θ ) ) , θ ( 0 , ) , I 0 + 2 α u ( 0 ) = 0 , lim θ D 0 + α 1 u ( θ ) = j = 1 m 2 β j D 0 + α 1 u ( ξ j ) ,
where α ( 1 , 2 ] , 0 < ξ 1 < ξ 2 < < ξ m 2 < , β j > 0 for j = 1 , , m 2 , D 0 + α is the Riemann–Liouville fractional derivative, I 0 + α is the Riemann–Liouville fractional integral, and j = 1 m 2 β j = 1 . By applying the Mawhin coincidence degree theory, they showed the existence of solutions for problem (5). In [7], the authors explored the Riemann–Liouville fractional differential equation with a p-Laplacian operator on the half-line, subject to integral boundary conditions at resonance as follows:
( φ p ( D 0 + α u ( θ ) ) ) + v ( t ) w ( θ , u ( θ ) , D 0 + α u ( θ ) ) = 0 , θ ( 0 , ) , φ p ( D 0 + α u ( 0 ) ) = 0 v ( θ ) φ p ( D 0 + α u ( θ ) ) d θ , φ p ( D 0 + α u ( ) ) = 0 v ( θ ) φ p ( D 0 + α u ( θ ) ) d θ ,
where α ( 0 , 1 ] , w : [ 0 , ) × R 2 R is a v-Carathéodory function; v L 1 [ 0 , ) ; v ( θ ) > 0 with 0 v ( θ ) d θ = 1 ; and φ p ( r ) = | r | p 2 r , where p > 1 . By using the Ge and Ren extension of the Mawhin coincidence degree theory, they proved the existence of solutions to problem (6). In [8], the authors studied the multipoint boundary value problem of the following Riemann–Liouville fractional p-Laplacian equation at resonance:
( φ p ( D 0 + α u ( θ ) ) ) + f ( θ , u ( θ ) , D 0 + α 1 u ( θ ) , D 0 + α u ( θ ) ) = 0 , θ ( 0 , ) , u ( 0 ) = u ( 0 ) = 0 , φ p ( D 0 + α u ( ) ) = j = 1 n α j φ p ( D 0 + α u ( ξ j ) ) ,
where α ( 1 , 2 ] , 0 < ξ 1 < ξ 2 < < ξ n < , α j > 0 , j = 1 n α j = 1 , and f is a L 1 -Carathéodory function. By using an extension of Mawhin continuation theorem, they showed the existence of solutions for problem (7). In [9], the authors examined the system of Riemann-Liouville fractional differential equations supplemented with the following boundary conditions:
D 0 + α u 1 ( θ ) = f 1 θ , u 1 ( θ ) , D 0 + β 1 u 2 ( θ ) , θ ( 0 , ) , D 0 + β u 2 ( θ ) = f 2 θ , u 2 ( θ ) , D 0 + α 1 u 1 ( θ ) , θ ( 0 , ) , u 1 ( 0 ) = 0 , lim θ D 0 + α 1 u 1 ( θ ) = j = 1 γ j u 1 ( η j ) , u 2 ( 0 ) = 0 , lim θ D 0 + β 1 u 2 ( θ ) = j = 1 σ j u 2 ( ξ j ) ,
where α , β ( 1 , 2 ] , 0 < η 1 < η 2 < η j < , 0 < ξ 1 < ξ 2 < < ξ j < , lim j η j = , lim j ξ j = , j = 1 | γ j | η j α < , j = 1 | σ j | ξ j β < , and f 1 , f 2 satisfy Carathéodory conditions. By use of the coincidence degree theory of Mawhin, they proved the existence of solutions of problem (8). We also mention recent papers [10,11,12,13] where the authors investigated various boundary value problems for Riemann–Liouville fractional differential equations on finite intervals at resonance. Different boundary value problems for Caputo fractional differential equations and systems on finite intervals at resonance were studied in [14,15,16,17].
Hadamard fractional calculus finds applications across diverse fields, including rheology and probability (see [18,19] and the references therein). For instance, extensions of the Lomnitz logarithmic creep law are used to model geophysical processes in fluid-like materials such as igneous rocks. For further recent developments concerning Hadamard fractional differential equations and systems with a variety of applications, we refer the reader to [20,21,22,23]. The novelty of the present problem (1), (2) lies in the treatment of a resonant boundary value problem for a Hadamard fractional differential equation of arbitrary order also containing varied fractional derivatives on an infinite interval. Moreover, the general boundary conditions considered here include Riemann–Stieltjes integrals and encompass a broad class of conditions such as finite or infinite multipoint boundary conditions, classical integral conditions, and combinations thereof.
The organization of our paper is outlined as follows. Section 2 provides several supporting lemmas used in the proofs of the main results. Section 3 is dedicated to presenting the existence theorems for problem (1), (2). Section 4 is devoted to two examples that demonstrates the applicability of the theoretical findings. Finally, Section 5 presents the conclusions of the paper.

2. Auxiliary Results

In this section, we provide the supporting results for our main theorems presented in Section 3.
We introduce the following expressions:
a = 1 A 1 s ( ln τ s ) 1 τ 3 d τ d P ( s ) = 1 A s 2 2 s 2 ln s 1 4 s 2 d P ( s ) , b = 1 A 1 s ( ln s τ ) 1 τ 3 ln τ d τ d P ( s ) = 1 A ln s + s 2 ln s + 1 s 2 4 s 2 d P ( s ) , c = 1 B s 1 τ 3 d τ d Q ( s ) = 1 B 1 2 s 2 d Q ( s ) , d = 1 B s 1 τ 3 ln τ d τ d Q ( s ) = 1 B 2 ln s + 1 4 s 2 d Q ( s ) , Δ = ad bc .
Firstly, we present our main assumptions:
(I1)
The functions P : [ 1 , A ] R and Q : [ 1 , B ] R are bounded variation functions satisfying the resonant conditions of 1 A d P ( σ ) = 1 ( P ( A ) P ( 1 ) = 1 ), 1 B d Q ( σ ) = 1 ( Q ( B ) Q ( 1 ) = 1 ), and 1 A ln σ d P ( σ ) = 0 . In addition, Δ 0 (given in (9)).
(I2)
The function f : ( 1 , ) × R n R is a Carathéodory function that satisfies the following conditions
(a)
f ( · , y 1 , y 2 , , y n ) is Lebesgue measurable for each ( y 1 , y 2 , , y n ) R n ;
(b)
f ( θ , · , · , , · ) is continuous (in the last n variables) for a.e. θ ( 1 , ) ;
(c)
There exist functions, i.e., p j L 1 ( 1 , ) , j = 1 , , n + 1 , and p j ( θ ) 0 , a.e. θ ( 1 , ) , j = 1 , , n + 1 such that
| f ( θ , y 1 , y 2 , , y n ) | p 1 ( θ ) | y 1 | 1 + ( ln θ ) α 1 + p 2 ( θ ) | y 2 | 1 + ( ln θ ) n 2 + + + p n 1 ( θ ) | y n 1 | 1 + ln θ + p n ( θ ) | y n | + p n + 1 ( θ )
for all y j R , j = 1 , , n and a.e. θ ( 1 , ) .
 Remark 1. 
If assumption ( I 1 ) is satisfied, then the boundary value problem (1), (2) is at resonance because the corresponding fractional differential equation, i.e.,
D 1 + α H u ( θ ) = 0 , θ ( 1 , ) ,
subject to the boundary conditions (2), admits nontrivial solutions of the form u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 for θ [ 1 , ) for arbitrary a 1 , a 2 R (see Lemma 2 below).
We define the Banach space of
U = u ; u , D 1 + α n + 1 H u , , D 1 + α 1 H u C [ 1 , ) , sup θ 1 | u ( θ ) | 1 + ( ln θ ) α 1 < , sup θ 1 | D 1 + α n + 1 H u ( θ ) | 1 + ( ln θ ) n 2 < , , sup θ 1 | D 1 + α 2 H u ( θ ) | 1 + ln θ < , sup θ 1 | D 1 + α 1 H u ( θ ) | < ,
with the following norm:
u U = max u α 1 , H D 1 + α n + 1 u n 2 , , H D 1 + α 2 u 1 , H D 1 + α 1 u 0 ,
where
u α 1 = sup θ 1 | u ( θ ) | 1 + ( ln θ ) α 1 , H D 1 + α n + 1 u n 2 = sup θ 1 | H D 1 + α n + 1 u ( θ ) | 1 + ( ln θ ) n 2 , , H D 1 + α 2 u 1 = sup θ 1 | H D 1 + α 2 u ( θ ) | 1 + ln θ , H D 1 + α 1 u 0 = sup θ 1 | H D 1 + α 1 u ( θ ) | , for u U .
We also define the Banach space of V = L 1 ( 1 , ) with a norm of v V = 1 | v ( θ ) | d θ .
We now define the linear operator as A : D ( A ) U V , given by
A u ( θ ) = H D 1 + α u ( θ ) , u D ( A ) , θ ( 1 , ) ,
where
D ( A ) = { u U ; D 1 + α H u V , a n d u verifies the conditions ( 2 ) } ,
and the nonlinear operator as B : U V , given by
B u ( θ ) = f ( θ , u ( θ ) , D 1 + α n + 1 H u ( θ ) , , D 1 + α 1 H u ( θ ) ) , θ ( 1 , ) , u U .
With the aid of the above operators ( A and B ), the problem (1), (2) can be written equivalently as the following operator equation:
A u = B u , u D ( A ) .
 Lemma 1 
([24]). Let K U be a bounded set satisfying the following conditions:
(a) The functions from K are equicontinuous on any compact interval of [ 1 , ) ;
(b) The functions from K are equiconvergent at infinity.
Therefore, set K is relatively compact in U .
 Lemma 2. 
We assume that ( I 1 ) holds. Then, the operator A : D ( A ) U V exhibits the following properties:
K e r ( A ) = { u D ( A ) ; u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 , θ [ 1 , ) , a 1 , a 2 R } U ,
I m A = { v V ; ( C 1 v ) ( θ ) = ( C 2 v ) ( θ ) = 0 , θ [ 1 , ) } V ,
where
( C 1 v ) ( θ ) = 1 A 1 σ ln σ τ v ( τ ) τ d τ d P ( σ ) , ( C 2 v ) ( θ ) = 1 B σ v ( τ ) τ d τ d Q ( σ ) ,
for all v V , θ [ 1 , ) .
 Proof. 
We consider u K e r A ; in other words, u D ( A ) and A u = 0 . Then, D 1 + α H u ( θ ) = 0 for all θ [ 1 , ) . Hence, according to Corollary 2.4 from [25], we find
u ( θ ) = a 1 ( ln θ ) α 1 + + a n ( ln θ ) α n , θ [ 1 , ) ,
with a k R , k = 1 , , n . By using the conditions of u ( k ) ( 1 ) = 0 , k = 0 , , n 3 , we infer that a k = 0 for k = 3 , , n . Therefore, according to (15), we obtain
u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 , θ [ 1 , ) .
Because D 1 + α 1 H u ( θ ) = a 1 Γ ( α ) and D 1 + α 2 H u ( θ ) = a 1 Γ ( α ) ln θ + a 2 Γ ( α 1 ) , according to ( I 1 ) , the following conditions are verified:
D 1 + α 2 H u ( 1 ) = 1 A D 1 + α 2 H u ( σ ) d P ( σ ) a 2 Γ ( α 1 ) = 1 A ( a 1 Γ ( α ) ln σ + a 2 Γ ( α 1 ) ) d P ( σ ) a 1 Γ ( α ) 1 A ln σ d P ( σ ) + a 2 Γ ( α 1 ) 1 A d P ( σ ) 1 = 0 , D 1 + α 1 H u ( ) = 1 B D 1 + α 1 H u ( σ ) d Q ( σ ) a 1 Γ ( α ) = 1 B a 1 Γ ( α ) d Q ( σ ) a 1 Γ ( α ) 1 B d Q ( σ ) 1 = 0 .
Therefore, we infer
K e r A { u D ( A ) , u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 , θ [ 1 , ) , a 1 , a 2 R } .
For the reverse inclusion, let u D ( A ) , u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 , θ [ 1 , ) with a 1 , a 2 R . We easily remark that D 1 + α H u ( θ ) = 0 for all θ [ 1 , ) , and u satisfies the conditions of (2), that is, u K e r A . So,
{ u D ( A ) ; u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 , θ [ 1 , ) , a 1 , a 2 R } K e r A .
Therefore, relation (12) holds. We see that d i m K e r A = 2 .
Now, we prove relation (13). We consider v I m A . Hence, there exists a function u D ( A ) such that A u = v or
D 1 + α H u ( θ ) = v ( θ ) , θ ( 1 , ) .
By applying the integral operator I 1 + α H to Equation (16), we find
u ( θ ) = I 1 + α H v ( θ ) + a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 + + a n ( ln θ ) α 1 ,
with a j R , j = 1 , , n . We use now the conditions of u ( j ) ( 1 ) = 0 , j = 0 , , n 3 , and we obtain a 3 = a 4 = = a n = 0 . Then, according to (17), we obtain
u ( θ ) = 1 Γ ( α ) 1 θ ln θ σ α 1 v ( σ ) σ d σ + a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 , θ [ 1 , ) .
Then,
D 1 + α 1 H u ( θ ) = I 1 + 1 H v ( θ ) + a 1 Γ ( α ) = 1 θ v ( σ ) σ d σ + a 1 Γ ( α ) , D 1 + α 2 H u ( θ ) = I 1 + 2 H v ( θ ) + a 1 Γ ( α ) ln θ + a 2 Γ ( α 1 ) = 1 θ ln θ σ v ( σ ) σ d σ + a 1 Γ ( α ) ln θ + a 2 Γ ( α 1 ) .
So, the conditions of D 1 + α 2 H u ( 1 ) = 1 A D 1 + α 2 H u ( σ ) d P ( σ ) and D 1 + α 1 H u ( ) = 1 B D 1 + α 1 H u ( σ ) d Q ( σ ) give us
a 2 Γ ( α 1 ) = 1 A 1 σ ln σ τ v ( τ ) τ d τ + a 1 Γ ( α ) ln σ + a 2 Γ ( α 1 ) d P ( σ ) , 1 v ( σ ) σ d σ + a 1 Γ ( α ) = 1 B 1 σ v ( τ ) τ d τ + a 1 Γ ( α ) d Q ( σ ) ,
or
a 1 Γ ( α ) 1 A ln σ d P ( σ ) a 2 Γ ( α 1 ) 1 A d P ( σ ) 1 = 1 A 1 σ ln σ τ v ( τ ) τ d τ d P ( σ ) , a 1 Γ ( α ) 1 1 B d Q ( σ ) = 1 B 1 σ v ( τ ) τ d τ d Q ( σ ) 1 v ( σ ) σ d σ .
According to the resonant conditions from ( I 1 ) , we find
1 A 1 σ ln σ τ v ( τ ) τ d τ d P ( σ ) = 0 , 1 B 1 σ v ( τ ) τ d τ d Q ( σ ) 1 v ( σ ) σ d σ = 0 .
The second relation from the above system is equivalent to
1 B 1 σ v ( τ ) τ d τ d Q ( σ ) = 1 B 1 v ( τ ) τ d τ d Q ( σ ) 1 B 1 v ( τ ) τ d τ 1 σ v ( τ ) τ d τ d Q ( σ ) = 0 1 B σ v ( τ ) τ d τ d Q ( σ ) = 0 .
Thus, ( C 1 v ) ( θ ) = ( C 2 v ) ( θ ) = 0 for all θ [ 1 , ) , that is,
I m A { v V ; ( C 1 v ) ( θ ) = ( C 2 v ) ( θ ) = 0 , θ [ 1 , ) } .
For the reverse inclusion, let v V with ( C 1 v ) ( θ ) = ( C 2 v ) ( θ ) = 0 for all θ [ 1 , ) . We consider u ( θ ) I 1 + α H v ( θ ) = 1 Γ ( α ) 1 θ ln θ σ α 1 v ( σ ) d σ σ , θ [ 1 , ) . It is evident that u D ( A ) . Indeed, we have u , D 1 + α n + 1 H u = I 1 + n 1 H u , , D 1 + α 1 H u = I 1 + 1 H u C [ 1 , ) , and D 1 + α H u = v V , with u U < . In addition, u verifies the conditions of (2), that is, u ( i ) ( 0 ) = 0 , i = 0 , , n 3 . Because
D 1 + α 2 H u ( θ ) = I 1 + 2 H v ( θ ) = 1 θ ln θ σ v ( σ ) d σ σ , D 1 + α 1 H u ( θ ) = I 1 + 1 H v ( θ ) = 1 θ v ( σ ) d σ σ ,
the conditions of
D 1 + α 2 H u ( 1 ) = 1 A D 1 + α 2 H u ( σ ) d P ( σ ) 0 = 1 A 1 σ ln σ τ v ( τ ) d τ τ d P ( σ )
and
D 1 + α 1 H u ( ) = 1 B D 1 + α 1 H u ( σ ) d Q ( σ ) 1 v ( σ ) d σ σ = 1 B 1 σ v ( τ ) d τ τ d Q ( σ ) 1 v ( σ ) σ d σ 1 B d Q ( σ ) = 1 B 1 σ v ( τ ) τ d τ d Q ( σ ) 1 B 1 v ( τ ) τ d τ d Q ( σ ) = 1 B 1 σ v ( τ ) τ d τ d Q ( σ ) 1 B 1 v ( τ ) τ d τ 1 σ v ( τ ) τ d τ d Q ( σ ) = 0 1 B σ v ( τ ) τ d τ d Q ( σ ) = 0 ,
are verified. We also find A u ( θ ) = v ( θ ) or D 1 + α H ( I 1 + α H v ( θ ) ) = v ( θ ) for all θ ( 1 , ) . Hence, we deduce
{ v V ; ( C 1 v ) ( θ ) = ( C 2 v ) ( θ ) = 0 , θ [ 1 , ) } I m A .
Then, according to (18) and (19), we derive relation (13). □
 Remark 2. 
The relation ( C 1 v ) ( θ ) = 0 can be equivalently written as
1 A τ A ln σ τ d P ( σ ) v ( τ ) τ d τ = 0 σ τ 1 A σ A ln τ σ d P ( τ ) v ( σ ) σ d σ = 0 ,
 and the relation ( C 2 v ) ( θ ) = 0 can be equivalently written as
1 B σ B v ( τ ) τ d τ d Q ( σ ) + 1 B B v ( τ ) τ d τ d Q ( σ ) = 0 1 B 1 τ d Q ( σ ) v ( τ ) τ d τ + B 1 B d Q ( σ ) v ( τ ) τ d τ = 0 1 B ( Q ( τ ) Q ( 1 ) ) v ( τ ) τ d τ + ( Q ( B ) Q ( 1 ) ) B v ( τ ) τ d τ = 0 τ = n o t σ 1 B ( Q ( σ ) Q ( 1 ) ) v ( σ ) σ d σ + B v ( σ ) σ d σ = 0 .
We now define the following linear operators:
( G 1 v ) ( θ ) = 1 Δ ( d C 1 v b C 2 v ) 1 θ 2 , ( G 2 v ) ( θ ) = 1 Δ ( a C 2 v c C 1 v ) 1 θ 2 , θ [ 1 , ) , v V .
 Lemma 3. 
Assume that ( I 1 ) holds. Then, the operator A : D ( A ) U V is a Fredholm operator of index zero. In addition, the operators E : U U and F : V V given by
E u ( θ ) = 1 Γ ( α ) D 1 + α 1 H u ( 1 ) ( ln θ ) α 1 + 1 Γ ( α 1 ) D + α 2 H u ( 1 ) ( ln θ ) α 2 , θ [ 1 , ) , u U , F v ( θ ) = G 1 v ( θ ) + ( G 2 v ( θ ) ) ln θ , θ [ 1 , ) , v V ,
 are linear projectors and verify the following relations:
I m E = K e r A , K e r F = I m A , U = K e r E K e r A , V = I m A I m F .
Proof. 
According to the definition of operator E , we deduce that it is a continuous and linear projector operator and I m E = K e r A (see also [3], Lemma 6). Furthermore, for u U , we have u = ( u E u ) + E u , that is, U = K e r E + K e r A . We can easily show that K e r E K e r A = { 0 } , so U = K e r E K e r A . We also observe that operator F is a continuous and linear operator and d i m I m F = 2 .
According to the definitions of G 1 and G 2 (see (22)), we deduce that for any v V and θ [ 1 , ) ,
G 1 ( G 1 v ) ( θ ) = 1 Δ ( d C 1 ( ( G 1 v ) ( θ ) ) b C 2 ( ( G 1 v ) ( θ ) ) ) 1 θ 2 = 1 Δ [ d 1 A 1 σ ln σ τ ( G 1 v ) ( τ ) τ d τ d P ( σ ) b 1 B σ ( G 1 v ) ( τ ) τ d τ d Q ( σ ) ] 1 θ 2 = 1 Δ [ d 1 A 1 σ ln σ τ 1 Δ ( d C 1 v b C 2 v ) 1 τ 3 d τ d P ( σ ) b 1 B σ 1 Δ ( d C 1 v b C 2 v ) 1 τ 3 d τ d Q ( σ ) ] 1 θ 2 = 1 Δ [ d Δ ( d C 1 v b C 2 v ) 1 A 1 σ ln τ σ 1 τ 3 d τ d P ( σ ) b Δ ( d C 1 v b C 2 v ) 1 B σ 1 τ 3 d τ d Q ( σ ) ] 1 θ 2 = 1 Δ · 1 Δ ( d C 1 v b C 2 v ) [ d 1 A 1 σ ln τ σ 1 τ 3 d τ d P ( σ ) b 1 B 1 2 σ 2 d Q ( σ ) ] 1 θ 2 = 1 Δ ( d C 1 v b C 2 v ) 1 θ 2 = ( G 1 v ) ( θ ) ,
G 2 ( ( G 2 v ) ( θ ) ln θ ) ( θ ) = 1 Δ ( a C 2 ( ( G 2 v ) ( θ ) ln θ ) c C 1 ( ( G 2 v ) ( θ ) ln θ ) ) 1 θ 2 = 1 Δ [ a 1 B σ ( G 2 v ) ( τ ) ln τ τ d τ d Q ( σ ) c 1 A 1 σ ln σ τ ( G 2 v ) ( τ ) ln τ τ d τ d P ( σ ) ] 1 θ 2 = 1 Δ [ a 1 B σ 1 Δ ( a C 2 v c C 1 v ) 1 τ 3 ln τ d τ d Q ( σ ) c 1 A 1 σ ln σ τ 1 Δ ( a C 2 v c C 1 v ) 1 τ 3 ln τ d τ d P ( σ ) ] 1 θ 2 = 1 Δ [ a Δ ( a C 2 v c C 1 v ) 1 B σ 1 τ 3 ln τ d τ d Q ( σ ) c Δ ( a C 2 v c C 1 v ) 1 A 1 σ ln σ τ 1 τ 3 ln τ d τ d P ( σ ) ] 1 θ 2 = 1 Δ · 1 Δ ( a C 2 v c C 1 v ) [ a 1 B σ 1 τ 3 ln τ d τ d Q ( σ ) c 1 A 1 σ ln σ τ 1 τ 3 ln τ d τ d P ( σ ) ] 1 θ 2 = 1 Δ ( a C 2 v c C 1 v ) 1 θ 2 = ( G 2 v ) ( θ ) ,
G 1 ( ( G 2 v ) ( θ ) ln θ ) ( θ ) = 1 Δ ( d C 1 ( ( G 2 v ) ( θ ) ln θ ) b C 2 ( ( G 2 v ) ( θ ) ln θ ) ) 1 θ 2 = 1 Δ [ d 1 A 1 σ ln σ τ ( G 2 v ) ( τ ) ln τ τ d τ d P ( σ ) b 1 B σ ( G 2 v ) ( τ ) ln τ τ d τ d Q ( σ ) ] 1 θ 2 = 1 Δ d 1 A 1 σ ln σ τ 1 Δ ( a C 2 v c C 1 v ) ln τ τ 3 d τ d P ( σ ) b 1 B σ 1 Δ ( a C 2 v c C 1 v ) ln τ τ 3 d τ d Q ( σ ) 1 θ 2 = 1 Δ · 1 Δ ( a C 2 v c C 1 v ) [ d 1 A 1 σ ln τ σ ln τ τ 3 d τ d P ( σ ) b 1 B ( σ ln τ τ 3 d τ ) d Q ( σ ) ] 1 θ 2 = 1 Δ · 1 Δ ( a C 2 v c C 1 v ) ( d b b d ) = 0 ,
G 2 ( ( G 1 v ) ( θ ) ) ( θ ) = 1 Δ ( a C 2 ( ( G 1 v ) ( θ ) ) c C 1 ( ( G 1 v ) ( θ ) ) ) 1 θ 2 = 1 Δ [ a 1 B σ ( G 1 v ) ( τ ) τ d τ d Q ( σ ) c 1 A 1 σ ln σ τ ( G 1 v ) ( τ ) τ d τ d P ( σ ) ] 1 θ 2 = 1 Δ [ a 1 B σ 1 Δ ( d C 1 v b C 2 v ) 1 τ 3 d τ d Q ( σ ) c 1 A 1 σ ln σ τ 1 Δ ( d C 1 v b C 2 v ) 1 τ 3 d τ d P ( σ ) ] 1 θ 2 = 1 Δ · 1 Δ ( d C 1 v b C 2 v ) [ 1 B σ 1 τ 3 d τ d Q ( σ ) c 1 A 1 σ ln σ τ 1 τ 3 d τ d P ( σ ) ] 1 θ 2 = 1 Δ · 1 Δ ( d C 1 v b C 1 v ) ( a c c a ) = 0 .
Then, based on the above relations, we deduce that F 2 v ( θ ) = F ( F v ) ( θ ) = F v ( θ ) for θ [ 1 , ) and v V , that is, F is a projector operator.
By using Lemma 2, we find I m A K e r F . In what follows, we show that K e r F I m A . To prove this, we consider v K e r F arbitrary but fixed for the moment. So, F v = 0 or G 1 v ( θ ) + ( G 2 v ( θ ) ) ln θ = 0 for all θ [ 1 , ) . Then, we infer
1 Δ ( d C 1 v b C 2 v ) 1 θ 2 + 1 Δ ( a C 2 v c C 1 v ) 1 θ 2 ln θ = 0 , θ [ 1 , ) .
Hence, we deduce the following homogeneous linear system in the unknowns of C 1 v and C 2 v :
d C 1 v b C 2 v = 0 , a C 2 v c C 1 v = 0 .
The determinant of the system (25) is Δ 0 , and we conclude that this system has a zero solution (unique solution) of C 1 v = 0 and C 2 v = 0 . This means that v I m A , so K e r F I m A . Then, we obtain K e r F = I m A .
With the aid of this last relation ( K e r F = I m A ), we find that V = I m A I m F . Clearly, given a function v V , we can formulate v = ( v F v ) + F v , so v F v K e r F = I m A . Hence, F v I m F . Hence, we derive v I m A + I m F . Furthermore, for any function v I m A I m F , we find C 1 v = C 2 v = 0 , and there exists w D ( F ) such that F w = v . Hence, there exist c 1 , c 2 R such that v ( θ ) = c 1 θ 2 + c 2 θ 2 ln θ , θ [ 1 , ) . So, we deduce the system as
1 A 1 σ ln σ τ c 1 τ 2 + c 2 τ 2 ln τ d τ τ d P ( σ ) = 0 , 1 B σ c 1 τ 2 + c 2 τ 2 ln τ d τ τ d Q ( σ ) = 0 ,
or
c 1 1 A 1 σ ln σ τ 1 τ 3 d τ d P ( σ ) + c 2 1 A 1 σ ln σ τ 1 τ 3 ln τ d τ d P ( σ ) = 0 , c 1 1 B σ 1 τ 3 d τ d Q ( σ ) + c 2 1 B σ 1 τ 3 ln τ d τ d Q ( σ ) = 0 .
Therefore we find the following system:
a c 1 + b c 2 = 0 , c c 1 + d c 2 = 0 ,
which has the unique solution c 1 = c 2 = 0 . Hence, we deduce that v = 0 , that is, I m A I m F = { 0 } . So, V = I m A I m F , and all the relations from (24) hold.
Then, according to (24), we deduce that dim K e r A = 2 and c o d i m I m A = 2 . Because I m A is closed, we infer that A is a Fredholm operator with the index I n d A = 0 . □
 Lemma 4. 
Suppose that ( I 1 ) holds. Then, the operator L E : I m A V D ( A ) K e r E U , given by
( L E v ) ( θ ) = I 1 + α H v ( θ ) = 1 Γ ( α ) 1 θ ln θ σ α 1 v ( σ ) d σ σ , v I m A ,
 represents the inverse of the operator A | D ( A ) K e r E . In addition, we have L E v U v V for all v I m A .
Proof. 
We show that L E is a well-defined operator on I m A . For v I m A (that is, v V , ( C 1 v ) ( θ ) = ( C 2 v ) ( θ ) = 0 for all θ [ 1 , ) ), we first prove that L E v D ( A ) .
Indeed, we have L E v U because
  • L E v C [ 1 , ) and
    sup θ 1 | L E v ( θ ) | 1 + ( ln θ ) α 1 sup θ 1 1 Γ ( α ) · 1 1 + ( ln θ ) α 1 1 θ ln θ σ α 1 | v ( σ ) | d σ σ = sup θ 1 1 Γ ( α ) 1 θ ( ln θ ln σ ) α 1 1 + ( ln θ ) α 1 | v ( σ ) | d σ σ sup θ 1 1 θ ( ln θ ) α 1 1 + ( ln θ ) α 1 | v ( σ ) | d σ σ sup θ 1 1 θ | v ( σ ) | d σ 1 | v ( σ ) | d σ = v V < ;
  • D 1 + α n + 1 H L E v = D 1 + α n + 1 H   I 1 + α H v = I 1 + n 1 H v C [ 1 , ) , and
    sup θ 1 | H D 1 + α n + 1 L E v ( θ ) | 1 + ( ln θ ) n 2 = sup θ 1 | H I 1 + n 1 v ( θ ) | 1 + ( ln θ ) n 2 sup θ 1 1 1 + ( ln θ ) n 2 1 Γ ( n 1 ) 1 θ ln θ σ n 2 | v ( σ ) | d σ σ = sup θ 1 1 Γ ( n 1 ) 1 θ ( ln θ ln σ ) n 2 1 + ( ln θ ) n 2 | v ( σ ) | d σ σ sup θ 1 1 θ ( ln θ ) n 2 1 + ( ln θ ) n 2 | v ( σ ) | d σ σ sup θ 1 1 θ | v ( σ ) | d σ 1 | v ( σ ) | d σ = v V < ;
  • D 1 + α 2 H L E v = D 1 + α 2 H   I 1 + α H v = I 1 + 2 H v C [ 1 , ) , and
    sup θ 1 | D 1 + α 2 H L E v ( θ ) | 1 + ln θ = sup θ 1 | I 1 + 2 H v ( θ ) | 1 + ln θ sup θ 1 1 1 + ln θ · 1 Γ ( 2 ) 1 θ ln θ σ | v ( σ ) | d σ σ = sup θ 1 1 Γ ( 2 ) 1 θ ln θ ln σ 1 + ln θ | v ( σ ) | d σ σ sup θ 1 1 θ ln θ 1 + ln θ | v ( σ ) | d σ σ sup θ 1 1 θ | v ( σ ) | d σ 1 | v ( σ ) | d σ = v V < ;
  • D 1 + α 1 H L E v = D 1 + α 1 H I 1 + α H v = I 1 + 1 H v C [ 1 , ) , and
    sup θ 1 | H D 1 + α 1 L E v ( θ ) | = sup θ 1 | H I 1 + 1 v ( θ ) | sup θ 1 1 θ ln θ σ 0 | v ( σ ) | d σ σ = sup θ 1 1 θ | v ( σ ) | d σ σ sup θ 1 1 θ | v ( σ ) | d σ 1 | v ( σ ) | d σ = v V < .
So, L E v U . In addition, we have D 1 + α H L E v = D 1 + α H I 1 + α H v = v V .
Additionally, L E v satisfies the conditions of (2). First, we find the following relations:
( L E v ) ( i ) ( 1 ) = D i   I 1 + α H v ( 1 ) = 0 , i = 0 , , n 3 ,
where D i is the classical derivative of order i. Indeed, for these relations, we obtain
  • For i = 0 , we have
    ( L E v ) ( 0 ) ( 1 ) = ( L E v ) ( 1 ) = D 1 + α H v ( 1 ) = 0 ;
  • For i = 1 , we find
    ( L E v ) ( 1 ) = d d θ ( L E v ) ( 1 ) = θ d d θ ( L E v ) ( 1 ) = δ ( L E v ) ( 1 ) = D 1 + 1 H ( L E v ) ( 1 ) = D 1 + 1 H I 1 + α H v ( 1 ) = I 1 + α 1 H v ( 1 ) = 0 ;
  • For i = 2 , we have
    ( L E v ) ( 1 ) = d 2 d θ 2 L E v ( θ ) | θ = 1 = d d θ L E v ( θ ) | θ = 1 + d 2 d θ 2 L E v ( θ ) | θ = 1 = θ d d θ L E v ( θ ) + θ 2 d 2 d θ 2 L E v ( θ ) | θ = 1 = θ d d θ L E v ( θ ) + θ d 2 d θ 2 L E v ( θ ) | θ = 1 = θ d d θ θ d d θ L E v ( θ ) | θ = 1 = θ d d θ 2 ( L E v ) ( 1 ) = δ 2 ( L E v ) ( 1 ) = D 1 + 2 H ( L E v ) ( 1 ) = H D 1 + 2 I 1 + α H v ( 1 ) = I 1 + α 2 H v ( 1 ) = 0 ;
  • For i = n 3 , we find
    ( L E v ) ( n 3 ) ( 1 ) = d n 3 d θ n 3 L E v ( θ ) | θ = 1 = = θ d d θ n 3 ( L E v ) ( 1 ) = δ n 3 ( L E v ) ( 1 ) = D 1 + n 3 H ( L E v ) ( 1 ) = D 1 + n 3 H   I 1 + α H v ( 1 ) = I 1 + α n + 3 H v ( 1 ) = 0 .
In fact, we can observe (as we proved before) that u ( i ) ( 1 ) = 0 for i = 0 , , n 3 if and only if D 1 + i H u ( 1 ) = ( δ i u ) ( 1 ) = ( θ d d θ ) i u ( θ ) | θ = 1 = 0 for i = 0 , , n 3 .
In addition, we obtain
D 1 + α 2 H ( L E v ) ( θ ) = D 1 + α 2 H ( I 1 + α H v ) ( θ ) = I 1 + 2 H v ( θ ) = 1 θ ln θ σ v ( σ ) d σ σ , D 1 + α 1 H ( L E v ) ( θ ) = D 1 + α 1 H ( I 1 + α H v ) ( θ ) = D 1 + 1 H v ( θ ) = 1 θ v ( σ ) d σ σ .
So, we deduce
D 1 + α 2 H ( L E v ) ( 1 ) = 0 , D 1 + α 1 H ( L E v ) ( ) = 1 v ( σ ) d σ σ .
Then, the last two conditions from (2) for L E v , namely
D 1 + α 2 H L E v ( 1 ) = 1 A D 1 + α 2 H L E v ( σ ) d P ( σ ) , a n d D 1 + α 1 H L E v ( ) = 1 B D 1 + α 1 H L E v ( σ ) d Q ( σ ) ,
can be written as
0 = 1 A 1 σ ln σ τ v ( τ ) d τ τ d P ( σ ) ,
and
1 v ( σ ) d σ σ = 1 B 1 σ v ( τ ) d τ τ d Q ( σ ) 1 B 1 σ v ( τ ) τ d τ d Q ( σ ) = 1 B 1 v ( τ ) τ d τ d Q ( σ ) 1 B 1 v ( τ ) τ d τ 1 σ v ( τ ) τ d τ d Q ( σ ) = 0 1 B σ v ( τ ) τ d τ d Q ( σ ) = 0 ,
respectively. The last relations are valid because ( C 1 v ) ( θ ) = ( C 2 v ) ( θ ) = 0 for all θ [ 1 , ) , ( v I m A ).
By a simple argument, we obtain L E v K e r E . Hence, we deduce that L E is a well-defined operator on I m A .
We now show that L E represents the inverse of the operator A | D ( A ) K e r E , that is,
 (a
A ( L E v ) ( θ ) = v ( θ ) , v I m A , θ ( 1 , ) ;
 (b
L E ( A u ) ( θ ) = u ( θ ) , u D ( A ) K e r E , θ ( 1 , ) .
Relation ( a ) is evident because
A ( L E v ) ( θ ) = H D 1 + α ( H I 1 + α v ) ( θ ) = v ( θ ) , θ ( 1 , ) , v I m A .
To prove relation ( b ) , let u D ( A ) K e r E . Then, according to Theorem 2.3 from [25], we have
L E ( A u ) ( θ ) = I 1 + α H ( H D 1 + α u ) ( θ ) = u ( θ ) + b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 + + b n ( ln θ ) α n ,
where b i R , i = 1 , , n . Since u D ( A ) , we obtain u ( j ) ( 1 ) = 0 for j = 0 , , n 3 and b j = 0 for j = 3 , , n . So, relation (27) gives us
L E ( A u ) ( θ ) = u ( θ ) + b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 , θ [ 1 , ) , b 1 , b 2 R .
Because u K e r E , we derive E u = 0 . So, according to (28), we find
E ( L E ( A u ) ) ( θ ) = E u ( θ ) + E ( b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 ) = E ( b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 ) = b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 , θ [ 1 , ) .
In addition, we have E ( L E ( A u ) ) ( θ ) = 0 for all θ [ 1 , ) due to L E ( w ) K e r E for all w I m A . Hence, according to (29), we obtain b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 = 0 for all θ [ 1 , ) . Therefore, relation (28) gives us L E ( A u ) ( θ ) = u ( θ ) for all θ [ 1 , ) .
Then, according to a ) and b ) , we deduce that L E = ( A | D ( A ) K e r E ) 1 .
Next, we prove that L E v U v V for all v I m A . For any v I m A , we saw before that
  • L E v α 1 = sup θ 1 | L E v ( θ ) | 1 + ( ln θ ) α 1 v V ,
  • H D 1 + α n + 1 L E v n 2 = sup θ 1 | H D 1 + α n + 1 L E v ( θ ) | 1 + ( ln θ ) n 2 v V ,
  • H D 1 + α 2 L E v 1 = sup θ 1 | H D 1 + α 2 L E v ( θ ) | 1 + ln θ v V ,
  • H D 1 + α 1 L E v 0 = sup θ 1 | H D 1 + α 1 L E v ( θ ) |   v V .
Therefore, L E v U v V for any v I m A .
 Lemma 5. 
Assume that ( I 1 ) and ( I 2 ) hold. Let Ω U be an open bounded set that satisfies the condition of D ( A ) Ω ¯ . Then, B is A -compact on Ω ¯ .
Proof. 
We use the characterization of the A -compact operator from [2], and we prove that
 (a
F B : Ω ¯ V is a continuous operator;
 (b
F B ( Ω ¯ ) is a bounded set;
 (c
L E , F B = L E ( I F ) B : Ω ¯ U is a continuous operator;
 (d
L E , F B ( Ω ¯ ) is a relatively compact set in U .
 (a
We obtain
F B ( u ) ( θ ) = G 1 B u ( θ ) + ( G 2 B u ( θ ) ) ln θ = 1 Δ d C 1 ( B u ) ( θ ) b C 2 ( B u ) ( θ ) 1 θ 2 + 1 Δ a C 2 ( B u ) ( θ ) c C 1 ( B u ) ( θ ) 1 θ 2 ln θ ,
where
C 1 ( B u ) ( θ ) = 1 A 1 σ ln σ τ B u ( τ ) τ d P ( σ ) = 1 A 1 σ ln σ τ f τ , u ( τ ) , H D 1 + α n + 1 u ( τ ) , , H D 1 + α 1 u ( τ ) d τ τ d P ( σ ) , C 2 ( B u ) ( θ ) = 1 B σ B u ( τ ) τ d τ d Q ( σ ) = 1 B σ f τ , u ( τ ) , H D 1 + α n + 1 u ( τ ) , , H D 1 + α 1 u ( τ ) d τ τ d Q ( σ ) ,
for all θ [ 1 , ) and u Ω ¯ . The f function is continuous in the last n variables; then, by way of the Lebesgue dominated convergence theorem, we infer that the C 1 B and C 2 B operators are continuous. We conclude that F B is a continuous operator.
 (b
First, we observe that there exists C > 0 such that u U C for all u Ω ¯ . Then, according to ( I 2 ) and (20), we obtain
| C 1 B u ( θ ) | = 1 A τ A ln σ τ d P ( σ ) B u ( τ ) τ d τ 1 A τ A ln σ τ d P ( σ ) · f τ , u ( τ ) , H D 1 + α n + 1 u ( τ ) , , H D 1 + α 1 u ( τ ) d τ τ 1 A τ A ln σ τ d P ( σ ) · p 1 ( τ ) | u ( τ ) | 1 + ( ln τ ) α 1 + p 2 ( τ ) | H D 1 + α n + 1 u ( τ ) | 1 + ( ln τ ) n 2 + + p n 1 ( τ ) | H D 1 + α 2 u ( τ ) | 1 + ln τ + p n ( τ ) | H D 1 + α 1 u ( τ ) | + p n + 1 ( τ ) d τ τ 1 A τ A ln σ τ d P ( σ ) max τ 1 | u ( τ ) 1 + ( ln τ ) α 1 p 1 ( τ ) + max τ 1 | H D 1 + α n + 1 u ( τ ) | 1 + ( ln τ ) n 2 p 2 ( τ ) + + max τ 1 | H D 1 + α 2 u ( τ ) | 1 + ln τ p n 1 ( τ ) + max τ 1 | H D 1 + α 1 u ( τ ) | p n ( τ ) + p n + 1 ( τ ) d τ τ .
Because
τ A ln σ τ d P ( σ ) sup σ [ τ , A ] ln σ τ V A τ ( P ) = ln A τ V A 1 ( P ) ( ln A ) V A 1 ( P ) ,
where V A 1 ( P ) is the total variation of function P on the interval of [ 1 , A ] , the above inequalities give us
| C 1 B u ( θ ) | ( ln A ) V A 1 ( P ) 1 u U ( p 1 ( τ ) + + p n ( τ ) ) + p n + 1 ( τ ) d τ ( ln A ) V A 1 ( P ) Ξ 0 u U + p n + 1 V : = M 1 ,
where Ξ 0 = i = 1 n p i V = i = 1 n 1 | p i ( τ ) | d τ and p n + 1 V = 1 | p n + 1 ( τ ) | d τ .
Next, according to ( I 2 ) and (21), we find
| C 2 B u ( θ ) | = 1 B ( Q ( σ ) Q ( 1 ) ) B u ( σ ) σ d σ + B B u ( σ ) σ d σ 2 sup σ [ 1 , B ] | Q ( σ ) | 1 B | B u ( σ ) | σ d σ + B | B u ( σ ) | σ d σ M 0 1 | B u ( σ ) | σ d σ M 0 1 f σ , u ( σ ) , H D 1 + α n + 1 u ( σ ) , , H D 1 + α 1 u ( σ ) d σ σ M 0 1 p 1 ( σ ) | u ( σ ) | 1 + ( ln σ ) α 1 + p 2 ( σ ) | H D 1 + α n + 1 u ( σ ) | 1 + ( ln σ ) n 2 + + p n 1 ( σ ) | H D 1 + α 2 u ( σ ) | 1 + ln σ + p n ( σ ) | H D 1 + α 1 u ( σ ) | + p n + 1 ( σ ) d σ M 0 1 p 1 ( σ ) max σ 1 | u ( σ ) | 1 + ( ln σ ) α 1 + p 2 ( σ ) max σ 1 | H D 1 + α n + 1 u ( σ ) | 1 + ( ln σ ) n 2 + + p n 1 ( σ ) max σ 1 | H D 1 + α 2 u ( σ ) | 1 + ln σ + p n ( σ ) max σ 1 | H D 1 + α 1 u ( σ ) | + p n + 1 ( σ ) d σ M 0 1 u U ( p 1 ( σ ) + p 2 ( σ ) + + p n ( σ ) ) + p n + 1 ( σ ) d σ M 0 Ξ 0 u U + p n + 1 V : = M 2 ,
where M 0 = max 2 sup σ [ 1 , B ] | Q ( σ ) | , 1 .
Hence, based on the above inequalities, we deduce
| F B u ( θ ) | 1 | Δ | ( | d | M 1 + | b | M 2 ) 1 θ 2 + 1 | Δ | ( | a | M 2 + | c | M 1 ) 1 θ 2 ln θ , θ [ 1 , ) , u Ω ¯ .
Therefore, we conclude
F B u V = 1 | F B u ( θ ) | d θ 1 1 | Δ | ( | d | M 1 + | b | M 2 ) 1 θ 2 d θ + 1 1 | Δ | ( | a | M 2 + | c | M 1 ) 1 θ 2 ln θ d θ = 1 | Δ | [ ( | d | + | c | ) M 1 + ( | b | + | a | ) M 2 ] : = M 3 , u Ω ¯ .
So, we infer that set F B ( Ω ¯ ) is bounded.
 (c
We remark that ( I F ) B is a continuous operator on Ω ¯ . In addition, we find
L E , F B ( u ) = L E ( I F ) B ( u ) = 1 Γ ( α ) 1 θ ln θ σ α 1 ( I F ) B u ( σ ) d σ σ = 1 Γ ( α ) 1 θ ln θ σ α 1 B u ( σ ) G 1 B u ( σ ) ( G 2 B u ( σ ) ) ln σ d σ σ = 1 Γ ( α ) 1 θ ln θ σ α 1 B u ( σ ) 1 Δ ( d C 1 ( B u ) b C 2 ( B u ) ) 1 σ 2 1 Δ ( a C 2 ( B u ) c C 1 ( B u ) ) 1 σ 2 ln σ d σ σ .
With the aid of the Lebesgue dominated convergence theorem for L E , F B ( u ) and its fractional derivatives, D 1 + α 1 H ( L E ( I F ) B ) ( u ) = I 1 + 1 H ( I F ) B ( u ) , D 1 + α 2 H ( L E ( I F ) B ) ( u ) =   I 1 + 2 H ( I F ) B ( u ) , …, and D 1 + α n + 1 H ( L E ( I F ) B ) ( u ) = I 1 + n 1 H ( I F ) B ( u ) , we conclude that L E , F B is a continuous operator in space U .
 (d
First, we prove that L E , F B : Ω ¯ U is bounded. In fact, for any u Ω ¯ , we have
B u V = 1 f σ , u ( σ ) , H D 1 + α n + 1 u ( σ ) , , H D 1 + α 1 u ( σ ) d σ Ξ 0 u U + p n + 1 V : = M 4 .
Then by use of Lemma 4, we find
L E ( I F ) B ( u ) U ( I F ) B ( u ) V B u V + F B u V M 4 + M 3 .
Then the functions L E ( I F ) B ( u ) and u Ω ¯ are uniformly bounded in U . Hence, L E ( I F ) B ( Ω ¯ ) is a bounded set in U .
Next, we prove that { L E ( I F ) B ( u ) , u Ω ¯ } are equicontinuous on any subcompact interval of [ 1 , ) . Indeed, for u Ω ¯ , according to ( I 2 ) , we find
| B u ( σ ) | = f ( σ , u ( σ ) , H D 1 + α n + 1 u ( σ ) , , H D 1 + α 1 u ( σ ) ) p 1 ( σ ) | u ( σ ) | 1 + ( ln σ ) α 1 + p 2 ( σ ) | H D 1 + α n + 1 u ( σ ) | 1 + ( ln σ ) n 2 + + p n 1 ( σ ) | H D 1 + α 2 u ( σ ) | 1 + ln σ + p n ( σ ) | H D 1 + α 1 u ( σ ) | + p n + 1 ( σ ) , a . e . σ ( 1 , ) .
Then, using similar inequalities as those obtained in part ( b ) , we deduce
| F B u ( σ ) | = | G 1 B u ( σ ) + ( G 2 B u ( σ ) ) ln σ | 1 | Δ | ( d C 1 B u ( σ ) b C 2 B u ( σ ) ) 1 σ 2 + 1 | Δ | ( a C 2 B u ( σ ) c C 1 B u ( σ ) ) ln σ σ 2 1 | Δ | ( | d | · | C 1 B u ( σ ) | + | b | · | C 2 B u ( σ ) | ) + ( | a | · | C 2 B u ( σ ) | + | c | · | C 1 B u ( σ ) | ) ln σ 1 σ 2 1 | Δ | [ ( | d | M 1 + | b | M 2 ) + ( | a | M 2 + | c | M 1 ) ln σ ] 1 σ 2 , σ [ 1 , ) ,
and
| ( I F ) B u ( σ ) | = | B u ( σ ) F B u ( σ ) | | B u ( σ ) | + | F B u ( σ ) | [ p 1 ( σ ) | u ( σ ) | 1 + ( ln σ ) α 1 + p 2 ( σ ) | H D 1 + α n + 1 u ( σ ) | 1 + ( ln σ ) n 2 + + p n 1 ( σ ) | H D 1 + α 2 u ( σ ) | 1 + ln σ + p n ( σ ) | H D 1 + α 1 u ( σ ) | + p n + 1 ( σ ) ] + 1 | Δ | [ ( | d | M 1 + | b | M 2 ) + ( | a | M 2 + | c | M 1 ) ln σ ] 1 σ 2 .
Now, let R be a finite positive constant from ( 1 , ) . Then, for any θ 1 , θ 2 [ 1 , R ] (we assume, without loss of generality, that θ 1 < θ 2 ), we obtain
L E , F B u ( θ 1 ) 1 + ( ln θ 1 ) α 1 L E , F B u ( θ 2 ) 1 + ( ln θ 2 ) α 1 = 1 Γ ( α ) 1 1 + ( ln θ 1 ) α 1 1 θ 1 ln θ 1 σ α 1 ( I F ) B u ( σ ) d σ σ 1 1 + ( ln θ 2 ) α 1 1 θ 2 ln θ 2 σ α 1 ( I F ) B u ( σ ) d σ σ 1 Γ ( α ) 1 θ 2 1 1 + ( ln θ 2 ) α 1 ln θ 2 σ α 1 ( I F ) B u ( σ ) d σ σ 1 θ 1 1 1 + ( ln θ 2 ) α 1 ln θ 2 σ α 1 ( I F ) B u ( σ ) d σ σ + 1 Γ ( α ) 1 θ 1 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 ( I F ) B u ( σ ) d σ σ 1 θ 1 ln θ 1 σ α 1 1 + ( ln θ 1 ) α 1 ( I F ) B u ( σ ) d σ σ = 1 Γ ( α ) θ 1 θ 2 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 ( I F ) B u ( σ ) d σ σ + 1 Γ ( α ) 1 θ 1 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 ln θ 1 σ α 1 1 + ( ln θ 1 ) α 1 ( I F ) B u ( σ ) d σ σ 1 Γ ( α ) θ 1 θ 2 | ( I F ) B u ( σ ) | d σ σ + 1 Γ ( α ) 1 θ 1 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 ln θ 1 σ α 1 1 + ( ln θ 1 ) α 1 × | ( I F ) B u ( σ ) | d σ 0 , as θ 1 θ 2 .
In a similar manner, we find
D 1 + α n + 1 H L E , F B u ( θ 1 ) 1 + ( ln θ 1 ) n 2 D 1 + α n + 1 H L E , F B u ( θ 2 ) 1 + ( ln θ 2 ) n 2 = I 1 + n 1 H ( I F ) B u ( θ 1 ) 1 + ( ln θ 1 ) n 2 I 1 + n 1 H ( I F ) B u ( θ 2 ) 1 + ( ln θ 2 ) n 2 0 , as θ 1 θ 2 , D 1 + α 2 H L E , F B u ( θ 1 ) 1 + ln θ 1 D 1 + α 2 H L E , F B u ( θ 2 ) 1 + ln θ 2 = I 1 + 2 H ( I F ) B u ( θ 1 ) 1 + ln θ 1 I 1 + 2 H ( I F ) B u ( θ 2 ) 1 + ln θ 2 0 , as θ 1 θ 2 , | H D 1 + α 1 L E , F B u ( θ 1 ) H D 1 + α 1 L E , F B u ( θ 2 ) | = | I 1 + 1 H ( I F ) B u ( θ 1 ) I 1 + 1 H ( I F ) B u ( θ 2 ) | θ 1 θ 2 | ( I F ) B u ( σ ) | d σ 0 , as θ 1 θ 2 .
Hence, we infer that { L E ( I F ) B ( u ) , u Ω ¯ } (or L E ( I F ) B ( Ω ¯ ) ) is a family of equicontinuous functions on [ 1 , R ] .
Finally, we show that L E , F B ( Ω ¯ ) is equiconvergent at infinity. For any u Ω ¯ , we have
1 | ( I F ) B u ( θ ) | d θ B u V + F B u V M 4 + M 3 .
Hence, for ε > 0 , there exists a constant T 1 > 1 such that
T 1 | ( I F ) B u ( θ ) | d θ < ε .
We also have
lim θ ln θ T 1 α 1 1 + ( ln θ ) α 1 = 1 , lim θ ln θ T 1 n 2 1 + ( ln θ ) n 2 = 1 , , lim θ ln θ T 1 1 + ln θ = 1 .
So, for ε > 0 above, there exists a constant T 2 > T 1 > 1 such that for any θ 1 , θ 2 T 2 and 1 σ T 1 , we obtain
ln θ 1 σ α 1 1 + ( ln θ 1 ) α 1 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 1 ln θ 1 T 1 α 1 1 + ( ln θ 1 ) α 1 + 1 ln θ 2 T 1 α 1 1 + ( ln θ 2 ) α 1 < ε , ln θ 1 σ n 2 1 + ( ln θ 1 ) n 2 ln θ 2 σ n 2 1 + ( ln θ 2 ) n 2 1 ln θ 1 T 1 n 2 1 + ( ln θ 1 ) n 2 + 1 ln θ 2 T 1 n 2 1 + ( ln θ 2 ) n 2 < ε , ln θ 1 σ 1 + ln θ 1 ln θ 2 σ 1 + ln θ 2 1 ln θ 1 T 1 1 + ln θ 1 + 1 ln θ 2 T 1 1 + ln θ 2 < ε .
Thus, for any θ 1 , θ 2 T 2 , according to (31) and (32), we deduce
L E , F B u ( θ 1 ) 1 + ( ln θ 1 ) α 1 L E , F B u ( θ 2 ) 1 + ( ln θ 2 ) α 1 = 1 Γ ( α ) 1 θ 1 ln θ 1 σ α 1 1 + ( ln θ 1 ) α 1 ( I F ) B u ( σ ) d σ σ 1 θ 2 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 ( I F ) B u ( σ ) d σ σ 1 Γ ( α ) 1 T 1 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 ln θ 1 σ α 1 1 + ( ln θ 1 ) α 1 · | ( I F ) B u ( σ ) | d σ σ + 1 Γ ( α ) T 1 θ 1 ln θ 1 σ α 1 1 + ( ln θ 1 ) α 1 | ( I F ) B u ( σ ) | d σ σ + 1 Γ ( α ) T 1 θ 2 ln θ 2 σ α 1 1 + ( ln θ 2 ) α 1 | ( I F ) B u ( σ ) | d σ σ ε Γ ( α ) 1 T 1 | ( I F ) B u ( σ ) | d σ σ + 2 Γ ( α ) T 1 | ( I F ) B u ( σ ) | d σ σ < ε Γ ( α ) ( M 4 + M 3 ) + 2 ε Γ ( α ) ε ( M 4 + M 3 + 2 ) .
In a similar manner, we obtain
D 1 + α n + 1 H L E , F B u ( θ 1 ) 1 + ( ln θ 1 ) n 2 D 1 + α n + 1 H L E , F B u ( θ 2 ) 1 + ( ln θ 2 ) n 2 = I 1 + n 1 H ( I F ) B u ( θ 1 ) 1 + ( ln θ 1 ) n 2 I 1 + n 1 H ( I F ) B u ( θ 2 ) 1 + ( ln θ 2 ) n 2 ε Γ ( n 1 ) ( M 4 + M 3 + 2 ) ε ( M 4 + M 3 + 2 ) , D 1 + α 2 H L E , F B u ( θ 1 ) 1 + ln θ 1 D 1 + α 2 H L E , F B u ( θ 2 ) 1 + ln θ 2 = I 1 + 2 H ( I F ) B u ( θ 1 ) 1 + ln θ 1 I 1 + 2 H ( I F ) B u ( θ 2 ) 1 + ln θ 2 ε Γ ( 2 ) ( M 4 + M 3 + 2 ) = ε ( M 4 + M 3 + 2 ) , | D 1 + α 1 H L E , F B u ( θ 1 ) D 1 + α 1 H L E , F B u ( θ 2 ) | = | I 1 + 1 H ( I F ) B u ( θ 1 ) I 1 + 1 H ( I F ) B u ( θ 2 ) | = θ 1 θ 2 ( I F ) B u ( σ ) d σ σ T 1 | ( I F ) B u ( σ ) | d σ < ε .
Thus, we deduce that L E , F B ( Ω ¯ ) is equiconvergent at infinity.
Therefore, by using Lemma 1, we conclude that L E , F B ( Ω ¯ ) is relatively compact in U . Hence, according to ( a ) ( d ) , we find that operator B is A -compact on Ω ¯ . □

3. Main Results

In this section, we state our main theorems concerning the existence of solutions to problem (1), (2).
 Theorem 1. 
Assume that ( I 1 ) and ( I 2 ) hold. We also suppose that
 (I3 )
There exist positive constants E > 0 and F > 1 such that for all u D ( A ) K e r A , if one of the following conditions is satisfied, then either C 1 ( B u ) 0 or C 2 ( B u ) 0 :
( i ) | H D 1 + α 2 u ( θ ) | > E for any θ [ 1 , F ] ;
( i i ) | H D 1 + α 1 u ( θ ) | > E for any θ [ 1 , ) .
 (I4) 
There exists a positive constant r such that for any b 1 , b 2 R satisfying | b 1 | > r or | b 2 | > r and u ( θ ) = b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 , then either
b 1 C 1 B ( u ( θ ) ) + b 2 C 2 B ( u ( θ ) ) < 0 ,
o r b 1 C 1 B ( u ( θ ) ) + b 2 C 2 B ( u ( θ ) ) > 0 .
 (I5) 
Ξ 0 ( 2 + Ξ 1 ln F ) < 1 ,
where Ξ 0 = i = 1 n p i V and Ξ 1 = max { 1 Γ ( α 1 ) , 1 } .
Then, the boundary value problem (1), (2) has at least one solution in space U .
The proof of Theorem 1 is based on the following three lemmas.
 Lemma 6. 
We assume that ( I 1 ) , ( I 2 ) , ( I 3 ) , and ( I 5 ) hold. Then, the set
Λ 1 = { u D ( A ) K e r A ; A u = ν B u , ν ( 0 , 1 ) } U
 is bounded in U .
Proof. 
For u Λ 1 , B u I m A , which implies C 1 ( B u ) ( θ ) = 0 and C 2 ( B u ) ( θ ) = 0 for all θ [ 1 , ) . Hence, according to assumption ( I 3 ) , we find that there exist θ 1 [ 1 , F ] and θ 2 [ 1 , ) such that | H D 1 + α 2 u ( θ 1 ) | E and | H D 1 + α 1 u ( θ 2 ) | E .
By using Lemma 4 and Remark 1 from [3], with α replaced by α 1 and α replaced by α 2 , we deduce
| D 1 + α 1 H u ( θ ) | = D 1 + α 1 H u ( θ 2 ) + θ 2 θ ( H D 1 + α u ) ( σ ) d σ σ D 1 + α 1 H u ( θ 2 ) + θ 2 θ D 1 + α H u ( σ ) d σ σ E + 1 | B u ( σ ) | d σ E + B u V , θ [ 1 , ) ,
and
| H D 1 + α 2 u ( θ ) | = D 1 + α 2 H u ( θ 1 ) + θ 1 θ ( H D 1 + α 1 u ) ( σ ) d σ σ | D 1 + α 2 H u ( θ 1 ) | + θ 1 θ | D 1 + α 1 H u ( σ ) | d σ σ E + H D 1 + α 1 u 0 θ 1 θ d σ σ E + D 1 + α 1 H u 0 ln θ θ 1 , θ [ 1 , ) .
So, we obtain
| D 1 + α 1 H u ( 1 ) | E + B u V , | D 1 + α 2 H u ( 1 ) | E + D 1 + α 1 H u 0 ln 1 θ 1 = E + D 1 + α 1 H u 0 ln θ 1 E + ln F D 1 + α 1 H u 0 E + ln F ( E + B u V ) = E ( 1 + ln F ) + ln F B u V .
Hence, according to the definition of E , we infer
E u α 1 = sup θ 1 | E u ( θ ) | 1 + ( ln θ ) α 1 = sup θ 1 1 1 + ( ln θ ) α 1 1 Γ ( α ) D 1 + α 1 H u ( 1 ) ( ln θ ) α 1 + 1 Γ ( α 1 ) D 1 + α 2 H u ( 1 ) ( ln θ ) α 2 1 Γ ( α ) | D 1 + α 1 H u ( 1 ) | sup θ 1 ( ln θ ) α 1 1 + ( ln θ ) α 1 + 1 Γ ( α 1 ) | D 1 + α 2 H u ( 1 ) | sup θ 1 ( ln θ ) α 2 1 + ( ln θ ) α 1 1 Γ ( α ) ( E + B u V ) + 1 Γ ( α 1 ) E ( 1 + ln F ) + ln F B u V .
Because D 1 + α 1 H E u ( θ ) = D 1 + α 1 H u ( 1 ) and D 1 + α 2 H E u ( θ ) = D 1 + α 1 H u ( 1 ) ln θ + D 1 + α 2 H u ( 1 ) , we obtain
D 1 + α 1 H E u 0 = sup θ 1 | D 1 + α 1 H E u ( θ ) | = sup θ 1 | D 1 + α 1 H u ( 1 ) | E + B u V , D 1 + α 2 H E u 1 = sup θ 1 | D 1 + α 2 H E u ( θ ) | 1 + ln θ = sup θ 1 | D 1 + α 1 H u ( 1 ) ln θ + D 1 + α 2 H u ( 1 ) | 1 + ln θ | D 1 + α 1 H u ( 1 ) | + | H D 1 + α 2 u ( 1 ) | E + B u V + E ( 1 + ln F ) + ln F B u V = 1 Γ ( 2 ) ( E + B u V ) + 1 Γ ( 1 ) [ E ( 1 + ln F ) + ln F B u V ] .
Next, we find
D 1 + α 3 H E u ( θ ) = 1 Γ ( 3 ) D 1 + α 1 H u ( 1 ) ( ln θ ) 2 + 1 Γ ( 2 ) D 1 + α 2 H u ( 1 ) ln θ , θ 1 ,
then
D 1 + α 3 H E u 2 = sup θ 1 | D 1 + α 3 H E u ( θ ) | 1 + ( ln θ ) 2 = sup θ 1 1 Γ ( 3 ) D 1 + α 1 H u ( 1 ) ( ln θ ) 2 + 1 Γ ( 2 ) D 1 + α 2 H u ( 1 ) ln θ 1 + ( ln θ ) 2 1 Γ ( 3 ) | D 1 + α 1 H u ( 1 ) | + 1 Γ ( 2 ) | D 1 + α 2 H u ( 1 ) | 1 Γ ( 3 ) ( E + B u V ) + 1 Γ ( 2 ) [ E ( 1 + ln F ) + ln F B u V ] .
We go on in this way until the fractional derivative of order α n + 1 of E u , and we obtain
D 1 + α n + 1 H E u ( θ ) = 1 Γ ( n 1 ) D 1 + α 1 H u ( 1 ) ( ln θ ) n 2 + 1 Γ ( n 2 ) D 1 + α 2 H u ( 1 ) ( ln θ ) n 3 , θ 1 ,
so
D 1 + α n + 1 H E u n 2 = sup θ 1 | D 1 + α n + 1 H E u ( θ ) | 1 + ( ln θ ) n 2 = sup θ 1 1 1 + ( ln θ ) n 2 1 Γ ( n 1 ) D 1 + α 1 H u ( 1 ) ( ln θ ) n 2 + 1 Γ ( n 2 ) D 1 + α 2 H u ( 1 ) ( ln θ ) n 3 1 Γ ( n 1 ) | D 1 + α 1 H u ( 1 ) | + 1 Γ ( n 2 ) | D 1 + α 2 H u ( 1 ) | 1 Γ ( n 1 ) ( E + B u V ) + 1 Γ ( n 2 ) [ E ( 1 + ln F ) + ln F B u V ] .
By use of the above inequalities, we derive
E u U = max E u α 1 , D 1 + α n + 1 H E u n 2 , , D 1 + α 3 H E u 2 , D 1 + α 2 H E u 1 , D 1 + α 1 H E u 0 = max 1 Γ ( α ) ( E + B u V ) + 1 Γ ( α 1 ) E ( 1 + ln F ) + ln F B u V , 1 Γ ( n 1 ) ( E + B u V ) + 1 Γ ( n 2 ) E ( 1 + ln F ) + ln F B u V , 1 Γ ( 3 ) ( E + B u V ) + 1 Γ ( 2 ) E ( 1 + ln F ) + ln F B u V , 1 Γ ( 2 ) ( E + B u V ) + 1 Γ ( 1 ) E ( 1 + ln F ) + ln F B u V , 1 Γ ( 1 ) E + B u V .
So, we obtain
E u U max 1 Γ ( α ) , 1 Γ ( n 1 ) , , 1 Γ ( 3 ) , 1 Γ ( 2 ) , 1 Γ ( 1 ) E + B u V + max 1 Γ ( α 1 ) , 1 Γ ( n 2 ) , , 1 Γ ( 2 ) , 1 Γ ( 1 ) E ( 1 + ln F ) + ln F B u V = max 1 Γ ( α ) , 1 E + B u V + max 1 Γ ( α 1 ) , 1 E ( 1 + ln F ) + ln F B u V = E + B u V + Ξ 1 E ( 1 + ln F ) + ln F B u V .
Therefore, we conclude
E u U E [ 1 + Ξ 1 ( 1 + ln F ) ] + ( 1 + Ξ 1 ln F ) B u V .
For u Λ 1 , we have ( I E ) u D ( A ) K e r E and A E u = 0 . Then, by virtue of Lemma 4, we find
( I E ) u U = L E A ( I E ) u U = L E A u U A u V ν B u V B u V .
Therefore, according to (36) and (37), we deduce
u U = E u + ( I E ) u U E u U + ( I E ) u U E [ 1 + Ξ 1 ( 1 + ln F ) ] + ( 1 + Ξ 1 ln F ) B u V + B u V = E [ 1 + Ξ 1 ( 1 + ln F ) ] + ( 2 + Ξ 1 ln F ) B u V .
According to ( I 2 ) , we obtain
B u V = 1 | B u ( θ ) | d θ = 1 f θ , u ( θ ) , H D 1 + α n + 1 u ( θ ) , , H D 1 + α 1 u ( θ ) d θ 1 ( p 1 ( θ ) | u ( θ ) | 1 + ( ln θ ) α 1 + p 2 ( θ ) | H D 1 + α n + 1 u ( θ ) | 1 + ( ln θ ) n 2 + + p n 1 ( θ ) | H D 1 + α 2 u ( θ ) | 1 + ln θ + p n ( θ ) | H D 1 + α 1 u ( θ ) | + p n + 1 ( θ ) ) d θ u α 1 p 1 V + H D 1 + α n + 1 u n 2 p 2 V + +   H D 1 + α 2 u 1 p n 1 V + H D 1 + α 1 u 0 p n V + p n + 1 V u U p 1 V + p 2 V + +   p n 1 V + p n V + p n + 1 V = Ξ 0 u U + p n + 1 V .
According to (38), (39), and (35), we find
u U E [ 1 + Ξ 1 ( 1 + ln F ) ] + ( 2 + Ξ 1 ln F ) Ξ 0 u U + p n + 1 V = E [ 1 + Ξ 1 ( 1 + ln F ) ] + ( 2 + Ξ 1 ln F ) Ξ 0 u U + ( 2 + Ξ 1 ln F ) p n + 1 V ,
then
u U E [ 1 + Ξ 1 ( 1 + ln F ) ] + ( 2 + Ξ 1 ln F ) p n + 1 V 1 ( 2 + Ξ 1 ln F ) Ξ 0 .
Thus, we deduce that Λ 1 is a bounded set in space U . □
 Lemma 7. 
Suppose that ( I 1 ) , ( I 2 ) , and ( I 4 ) hold. Then, the set
Λ 2 = { u K e r A ; B u I m A } U ,
 is bounded in U .
Proof. 
Let u Λ 2 ; then, u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 , with a 1 , a 2 R and C 1 ( B u ) ( θ ) = C 2 ( B u ) ( θ ) = 0 for all θ [ 1 , ) . From ( I 4 ) , we deduce that | a 1 | r and | a 2 | r . Hence, we have
u α 1 = sup θ 1 | u ( θ ) | 1 + ( ln θ ) α 1 = sup θ 1 | a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 | 1 + ( ln θ ) α 1 | a 1 | + | a 2 | 2 r , H D 1 + α 1 u 0 = sup θ 1 | H D 1 + α 1 u ( θ ) | = sup θ 1 | a 1 Γ ( α ) | =   | a 1 Γ ( α ) | r Γ ( α ) , H D 1 + α 2 u 1 = sup θ 1 | H D 1 + α 2 u ( θ ) | 1 + ln θ = sup θ 1 | a 1 Γ ( α ) ln θ + a 2 Γ ( α 1 ) | 1 + ln θ | a 1 Γ ( α ) | + | a 2 Γ ( α 1 ) | r ( Γ ( α ) + Γ ( α 1 ) ) , H D 1 + α 3 u 2 = sup θ 1 | H D 1 + α 3 u ( θ ) | 1 + ( ln θ ) 2 = sup θ 1 a 1 Γ ( α ) Γ ( 3 ) ( ln θ ) 2 + a 2 Γ ( α 1 ) Γ ( 2 ) ln θ 1 + ( ln θ ) 2 a 1 Γ ( α ) Γ ( 3 ) + a 2 Γ ( α 1 ) Γ ( 2 ) r Γ ( α ) Γ ( 3 ) + Γ ( α 1 ) Γ ( 2 ) = r Γ ( α ) 2 + Γ ( α 1 ) .
We proceed with these estimations up to the fractional derivative of order α n + 1 of function u , and we infer
H D 1 + α n + 1 u n 2 = sup θ 1 | H D 1 + α n + 1 u ( θ ) | 1 + ( ln θ ) n 2 = sup θ 1 a 1 Γ ( α ) Γ ( n 1 ) ( ln θ ) n 2 + a 2 Γ ( α 1 ) Γ ( n 2 ) ( ln θ ) n 3 1 + ( ln θ ) n 2 a 1 Γ ( α ) Γ ( n 1 ) + a 2 Γ ( α 1 ) Γ ( n 2 ) r Γ ( α ) Γ ( n 1 ) + Γ ( α 1 ) Γ ( n 2 ) .
With the help of the above inequalities, we deduce
u U = max { 2 r , r Γ ( α ) Γ ( n 1 ) + Γ ( α 1 ) Γ ( n 2 ) , , r Γ ( α ) 2 + Γ ( α 1 ) , r ( Γ ( α ) + Γ ( α 1 ) ) , r Γ ( α ) } < .
We conclude that Λ 2 is a bounded set in space U . □
 Lemma 8. 
Suppose that ( I 1 ) , ( I 2 ) , and ( I 4 ) hold. Then the set
Λ 3 = { u K e r A , ς ζ H u + ( 1 ζ ) F B u = 0 , ζ [ 0 , 1 ] } U
 is bounded in U , where ς = 1 if relation (33) holds and ς = 1 if relation (34) holds. The operator H : K e r A I m F is a linear isomorphism given by
H ( u ) ( θ ) = 1 Δ ( d b 1 b b 2 ) 1 θ 2 + ( a b 2 c b 1 ) ln θ θ 2 , θ [ 1 , ) ,
 for u K e r A , u ( θ ) = b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 , θ 1 , with b 1 , b 2 R .
Proof. 
We assume that assumption (34) holds, that is, ς = 1 . For u Λ 3 , we can write u in the form of u ( θ ) = b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 , θ 1 , and ζ H u = ( 1 ζ ) F B u for ζ [ 0 , 1 ] . Using the same argument as in the proof of Lemma 7, we need only to prove that | b 1 | r and | b 2 | r . We have three cases:
( 1 ) For ζ = 0 , F B u = 0 , that is, G 1 ( B u ) ( θ ) + ( G 2 ( B u ) ( θ ) ) ln θ = 0 for all θ [ 1 , ) , or
1 Δ ( d C 1 B u b C 2 B u ) 1 θ 2 + 1 Δ ( a C 2 B u c C 1 B u ) ln θ θ 2 = 0 , θ [ 1 , ) .
Then, we find
d C 1 B u b C 2 B u = 0 , c C 1 B u + a C 2 B u = 0 .
Since the determinant of the above system is Δ 0 , we deduce that C 1 ( B u ) = C 2 ( B u ) = 0 . Hence, according to ( I 4 ) , we obtain | b 1 | r and | b 2 | r .
( 2 ) For ζ = 1 , H u = 0 , that is.
1 Δ ( d b 1 b b 2 ) 1 θ 2 + ( a b 2 c b 1 ) ln θ θ 2 = 0 .
From this relation, it follows that
d b 1 b b 2 = 0 , a b 2 c b 1 = 0 .
Since Δ 0 , we obtain b 1 = b 2 = 0 .
( 3 ) For ζ ( 0 , 1 ) , based on ζ H u = ( 1 ζ ) F B u , we have
ζ 1 Δ ( d b 1 b b 2 ) 1 θ 2 + 1 Δ ( a b 2 c b 1 ) ln θ θ 2 = ( 1 ζ ) 1 Δ ( d C 1 B u b C 2 B u ) 1 θ 2 + 1 Δ ( a C 2 B u c C 1 B u ) ln θ θ 2 ,
from which we deduce
ζ d b 1 ζ b b 2 = ( 1 ζ ) d C 1 B u + ( 1 ζ ) b C 2 B u , ζ a b 2 ζ c b 1 = ( 1 ζ ) a C 2 B u + ( 1 ζ ) c C 1 B u .
Because Δ 0 , we find
ζ b 1 = ( 1 ζ ) C 1 B u , ζ b 2 = ( 1 ζ ) C 2 B u .
Now, we show that | b 1 | r and | b 2 | r . If this assertion were not to hold, then according to (34), we would obtain
0 ζ ( b 1 2 + b 2 2 ) = ( ζ b 1 ) b 1 + ( ζ b 2 ) b 2 = b 1 ( 1 ζ ) C 1 B u b 2 ( 1 ζ ) C 2 B u = ( 1 ζ ) ( b 1 C 1 B u + b 2 C 2 B u ) < 0 .
This is a contradiction. Consequently, we derive | b 1 | r and | b 2 | r . We infer that Λ 3 is a bounded set in U . If relation (33) holds (for ζ = 1 ), then by way of similar arguments, we find that Λ 3 is also bounded. □
We now present the proof of Theorem 1.
 Proof of Theorem 1. 
Let Λ 0 U be a bounded open set such that i = 1 3 Λ ¯ i Λ 0 . In Lemma 3, we saw that A is a Fredholm operator of index 0, and in Lemma 5, we found that B is an A -compact operator on Λ ¯ 0 . Furthermore, according to Lemmas 6 and 7, we have
 (i
A u ν B u for any u ( D ( A ) K e r A ) Λ 0 , ν ( 0 , 1 ) ;
 (ii
B u I m A for any u K e r A Λ 0 .
Next, we prove that condition ( i i i ) from the Mawhin continuation theorem (see Theorem IV.13 from [1] or Theorem 1 from [3]) is also satisfied. For this, we define the operator S ( u , ζ ) = ς ζ H u + ( 1 ζ ) F B u , where ς is given in Lemma 8. So, by way of Lemma 8, we find S ( u , ζ ) 0 for all u K e r A Λ 0 and ζ [ 0 , 1 ] . Hence, by the homotopy of degree, we conclude
d e g { F B | K e r A , Λ 0 K e r A , 0 } = d e g { S ( · , 0 ) , Λ 0 K e r A , 0 } = d e g { S ( · , 1 ) , Λ 0 K e r A , 0 } = d e g { ς H , Λ 0 K e r A , 0 } 0 .
Then, all assumptions of Theorem IV.13 from [1] ( ( i ) , ( i i ) , and ( i i i ) ) are satisfied. Therefore, we deduce that the operator Equation (11)—specifically, A u = B u —admits at least one solution u D ( A ) Λ ¯ 0 . This implies that problem (1), (2) has at least one solution in space U . □
 Theorem 2. 
Suppose that ( I 1 ) and ( I 2 ) hold. We also suppose that
 (I6)
There exists a positive constant G such that for each u D ( A ) satisfying | H D 1 + α 1 u ( θ ) | > G for all θ [ 1 , ) , we have either
s g n { H D 1 + α 1 u ( θ ) } C 2 B u ( θ ) > 0 , θ [ 1 , ) ,
or s g n { H D 1 + α 1 u ( θ ) } C 2 B u ( θ ) < 0 , θ [ 1 , ) .
 (I7)
There exist positive constants L > 0 and M > 1 such that for every u D ( A ) satisfying | H D 1 + α 2 u ( θ ) | > L for all θ [ 1 , M ] , we have either
s g n { H D 1 + α 2 u ( θ ) } C 1 B u ( θ ) > 0 , θ [ 1 , M ] ,
or s g n { H D 1 + α 2 u ( θ ) } C 1 B u ( θ ) > 0 , θ [ 1 , M ] .
 (I8)
Ξ 0 Ξ 2 < 1 ,
where Ξ 0 = i = 1 n p i V , Ξ 2 = max { 3 Γ ( α ) , 3 2 } + 2 Ξ 1 ln M , with Ξ 1 = max { 1 Γ ( α 1 ) , 1 } .
Then, the boundary value problem (1), (2) has at least one solution in space U .
The proof of Theorem 2 is based on the following three lemmas.
 Lemma 9. 
Assume that ( I 1 ) , ( I 2 ) , ( I 6 ) , ( I 7 ) , and ( I 8 ) hold. Then, the set Λ 1 (defined in Lemma 6) is bounded in U .
Proof. 
We consider u Λ 1 ; then, we have B u I m A = K e r A . By virtue of ( I 6 ) and ( I 7 ) , there exist constants θ 1 [ 1 , ) and θ 2 [ 1 , M ] such that | H D 1 + α 1 u ( θ 1 ) | G and | H D 1 + α 2 u ( θ 2 ) | L . This, together with Lemma 4 and Remark 1 from [3], gives us
D 1 + α 1 H u ( θ ) = D 1 + α 1 H u ( θ 1 ) + θ 1 θ D 1 + α H u ( σ ) d σ σ , D 1 + α 2 H u ( θ ) = D 1 + α 2 H u ( θ 2 ) + θ 2 θ D 1 α 1 H u ( σ ) d σ σ = D 1 + α 2 H u ( θ 2 ) + ( ln θ ln θ 2 ) D 1 + α 1 H u ( θ 1 ) + θ 2 θ θ 1 σ D 1 + α H u ( τ ) d τ τ d σ σ .
Then, we obtain
D 1 + α 1 H u 0 = sup θ 1 | D 1 + α 1 H u ( θ ) | G + D 1 + α H u V ,
and
D 1 + α 2 H u 1 = sup θ 1 | D 1 + α 2 H u ( θ ) | 1 + ln θ sup θ 1 | D 1 + α 2 H u ( θ 2 ) | 1 + ln θ + sup θ 1 | ln θ ln θ 2 | 1 + ln θ | D 1 + α 1 H u ( θ 1 ) | + sup θ 1 1 1 + ln θ θ 2 θ θ 1 σ D 1 + α H u ( τ ) d τ τ d σ σ L + G + 1 | D 1 + α H u ( τ ) d τ sup θ 1 1 1 + ln θ θ 2 θ d σ σ = L + G + sup θ 1 | ln θ ln θ 2 | 1 + ln θ D 1 + α H u V = L + G + D 1 + α H u V .
In addition, for u Λ 1 D ( A ) , we find
u ( θ ) = I 1 + α H D 1 + α H u ( θ ) + c 1 ( ln θ ) α 1 + c 2 ( ln θ ) α 2 + + c n ( ln θ ) α n , θ 1 ,
with c i R , i = 1 , , n . Because u satisfies the boundary conditions of (2), we obtain
u ( θ ) = I 1 + α H D 1 + α H u ( θ ) + c 1 ( ln θ ) α 1 + c 2 ( ln θ ) α 2 , θ 1 .
Therefore, we find
u ( θ ) 1 + ( ln θ ) α 1 = 1 Γ ( α ) 1 θ ( ln θ σ ) α 1 1 + ( ln θ ) α 1 D 1 + α H u ( σ ) d σ + c 1 ( ln θ ) α 1 1 + ( ln θ ) α 1 + c 2 ( ln θ ) α 2 1 + ( ln θ ) α 1 ,
D 1 + α 1 H u ( θ ) = D 1 + α 1 H I 1 + α H D 1 + α H u ( θ ) + c 1 D 1 + α 1 H ( ( ln θ ) α 1 ) + c 2 D 1 + α 1 H ( ( ln θ ) α 2 ) = I 1 + 1 H D 1 + α H u ( θ ) + c 1 Γ ( α ) = 1 θ D 1 + α H u ( σ ) d σ σ + c 1 Γ ( α ) ,
and
D 1 + α 2 H u ( θ ) = D 1 + α 2 H I 1 + α H D 1 + α H u ( θ ) + c 1 D 1 + α 2 H ( ( ln θ ) α 1 ) + c 2 D 1 + α 2 H ( ( ln θ ) α 2 ) = I 1 + 2 H D 1 + α H u ( θ ) + c 1 Γ ( α ) Γ ( 2 ) ln θ + c 2 Γ ( α 1 ) Γ ( 1 ) = 1 θ ln θ σ D 1 + α H u ( σ ) d σ σ + c 1 Γ ( α ) ln θ + c 2 Γ ( α 1 ) = 1 θ ln σ D 1 + α H u ( σ ) d σ σ + ln θ 1 θ D 1 + α H u ( σ ) d σ σ + c 1 Γ ( α ) ln θ + c 2 Γ ( α 1 ) = 1 θ ln σ D 1 + α H u ( σ ) d σ σ + ln θ D 1 + α 1 H u ( θ ) c 1 Γ ( α ) + c 1 Γ ( α ) ln θ + c 2 Γ ( α 1 ) = 1 θ ln σ D 1 + α H u ( σ ) d σ σ + ln θ D 1 + α 1 H u ( θ ) + c 2 Γ ( α 1 ) .
So, for θ = θ 2 , relation (49) gives us
D 1 + α 2 H u ( θ 2 ) = 1 θ 2 ln σ D 1 + α H u ( σ ) d σ σ + ln θ 2 D 1 + α 1 H u ( θ 2 ) + c 2 Γ ( α 1 ) .
Hence, from (48) and (50), we obtain the following relations for constants c 1 and c 2 :
c 1 = 1 Γ ( α ) D 1 + α 1 H u ( θ ) 1 θ D 1 + α H u ( σ ) d σ σ , c 2 = 1 Γ ( α 1 ) D 1 + α 2 H u ( θ 2 ) ln θ 2 D 1 + α 1 H u ( θ 2 ) + 1 θ 2 ln σ D 1 + α H u ( σ ) d σ σ .
Now, by using the relations (45), (46), and (51), we derive
| c 1 |     1 Γ ( α ) D 1 + α 1 H u 0 + D 1 + α H u V 1 Γ ( α ) G + 2 D 1 + α H u V , | c 2 |     1 Γ ( α 1 ) L + ln M D 1 + α 1 H u 0 + ln M D 1 + α H u V 1 Γ ( α 1 ) L + G ln M + 2 ln M D 1 + α H u V .
We substitute (52) in (47), and we obtain
u ( θ ) 1 + ( ln θ ) α 1 1 Γ ( α ) D 1 + α H u V   +   | c 1 |   +   | c 2 | 1 Γ ( α ) D 1 + α H u V + 1 Γ ( α ) G + 2 D 1 + α H u V + 1 Γ ( α 1 ) L + G ln M + 2 ln M D 1 + α H u V = 1 Γ ( α ) D 1 + α H u V [ 3 + 2 ( α 1 ) ln M ] + G Γ ( α ) + L + G ln M Γ ( α 1 ) , θ [ 1 , ) .
From this last inequality, we find
u α 1 1 Γ ( α ) [ 3 + 2 ( α 1 ) ln M ] D 1 + α H u V + G Γ ( α ) + L + G ln M Γ ( α 1 ) .
We continue with the fractional derivative of order α 3 of function u , and we obtain
D 1 + α 3 H u ( θ ) = D 1 + α 3 H D 1 + α H D 1 + α H u ( θ ) + c 1 D 1 + α 3 H ( ( ln θ ) α 1 ) + c 2 D 1 + α 3 H ( ( ln θ ) α 2 ) = I 1 + 3 H D 1 + 1 H u ( θ ) + c 1 Γ ( α ) Γ ( 3 ) ( ln θ ) 2 + c 2 Γ ( α 1 ) Γ ( 2 ) ln θ = 1 Γ ( 3 ) 1 θ ln θ σ 2 D 1 + α H u ( σ ) d σ σ + c 1 Γ ( α ) Γ ( 3 ) ( ln θ ) 2 + c 2 Γ ( α 1 ) Γ ( 2 ) ln θ ,
and
D 1 + α 3 H u 2 = sup θ 1 | D 1 + α 3 H u ( θ ) | 1 + ( ln θ ) 2 sup θ 1 1 Γ ( 3 ) 1 θ ( ln θ σ ) 2 1 + ( ln θ ) 2 | D 1 + α H u ( σ ) | d σ σ +   | c 1 | Γ ( α ) Γ ( 3 ) sup θ 1 ( ln θ ) 2 1 + ( ln θ ) 2 + | c 2 | Γ ( α 1 ) Γ ( 2 ) sup θ 1 ln θ 1 + ( ln θ ) 2 1 Γ ( 3 ) D 1 + α H u V + Γ ( α ) Γ ( 3 ) | c 1 | + Γ ( α 1 ) Γ ( 2 ) | c 2 | 1 Γ ( 3 ) D 1 + α H u V + 1 Γ ( 3 ) ( G + 2 D 1 + α H u V ) + 1 Γ ( 2 ) ( L + G ln M + 2 ln M D 1 + α H u V ) .
We proceed with these estimations until the fractional derivative of order α n + 1 of function u , and we infer
D 1 + α n + 1 H u ( θ ) = D 1 + α n + 1 H I 1 + α H D 1 + α H u ( θ ) + c 1 D 1 + α n + 1 H ( ( ln θ ) α 1 ) + c 2 D 1 + α n + 1 H ( ( ln θ ) α 2 ) = I 1 + n 1 H D 1 + α H u ( θ ) + c 1 Γ ( α ) Γ ( n 1 ) ( ln θ ) n 2 + c 2 Γ ( α 1 ) Γ ( n 2 ) ( ln θ ) n 3 = 1 Γ ( n 1 ) 1 θ ln θ σ n 2 D 1 + α H u ( σ ) d σ σ + c 1 Γ ( α ) Γ ( n 1 ) ( ln θ ) n 2 + c 2 Γ ( α 1 ) Γ ( n 2 ) ( ln θ ) n 3 ,
and
D 1 + α n + 1 H u n 2 = sup θ 1 | D 1 + α n + 1 H u ( θ ) | 1 + ( ln θ ) n 2 1 Γ ( n 1 ) sup θ 1 1 θ ln θ σ n 2 1 + ( ln θ ) n 2 | D 1 + α H u ( σ ) | d σ σ +   | c 1 | Γ ( α ) Γ ( n 1 ) sup θ 1 ( ln θ ) n 2 1 + ( ln θ ) n 2 + | c 2 | Γ ( α 1 ) Γ ( n 2 ) sup θ 1 ( ln θ ) n 3 1 + ( ln θ ) n 2 1 Γ ( n 1 ) D 1 + α H u V + 1 Γ ( n 1 ) ( G + 2 D 1 + α H u V ) + 1 Γ ( n 2 ) ( L + G ln M + 2 ln M D 1 + α H u V ) .
Combining relations (45), (46) and (53)–(55), we deduce
u U = max 1 Γ ( α ) [ 3 + 2 ( α 1 ) ln M ] D 1 + α H u V + G Γ ( α ) + L + G ln M Γ ( α 1 ) , 1 Γ ( n 1 ) + 2 Γ ( n 1 ) + 2 ln M Γ ( n 2 ) D 1 + α H u V + G Γ ( n 1 ) + L + G ln M Γ ( n 2 ) , , 1 Γ ( 3 ) + 2 Γ ( 3 ) + 2 ln M Γ ( 2 ) D 1 + α H u V + G Γ ( 3 ) + L + G ln M Γ ( 2 ) , D 1 + α H u V + L + G , D 1 + α H u V + G = max 1 Γ ( α ) [ 3 + 2 ( α 1 ) ln M ] D 1 + α H u V + G Γ ( α ) + L + G ln M Γ ( α 1 ) , 1 Γ ( n 1 ) [ 3 + 2 ( n 2 ) ln M ] D 1 + α H u V + G Γ ( n 1 ) + L + G ln M Γ ( n 2 ) , , 1 Γ ( 3 ) [ 3 + 2 · 2 ln M ] D 1 + α H u V + G Γ ( 3 ) + L + G ln M Γ ( 2 ) , D 1 + α H u V + L + G , D 1 + α H u V + G } max 1 Γ ( α ) [ 3 + 2 ( α 1 ) ln M ] , 1 Γ ( n 1 ) [ 3 + 2 ( n 2 ) ln M ] , , 1 Γ ( 3 ) [ 3 + 2 · 2 ln M ] , 1 D 1 + α H u V + G max 1 Γ ( α ) , 1 Γ ( n 1 ) , , 1 Γ ( 3 ) , 1 Γ ( 2 ) , 1 Γ ( 1 ) + L max 1 Γ ( α 1 ) , 1 Γ ( n 2 ) , , 1 Γ ( 2 ) , 1 + G ln M max 1 Γ ( α 1 ) , 1 Γ ( n 2 ) , , 1 Γ ( 2 ) .
Because max 1 Γ ( α ) , 1 Γ ( n 1 ) , , 1 Γ ( 2 ) , 1 Γ ( 1 ) = max 1 Γ ( α ) , 1 = 1 and
max 1 Γ ( α ) [ 3 + 2 ( α 1 ) ln M ] , 1 Γ ( n 1 ) [ 3 + 2 ( n 2 ) ln M ] , , 1 Γ ( 3 ) [ 3 + 2 · 2 ln M ] , 1 max 3 Γ ( α ) , 3 Γ ( n 1 ) , , 3 Γ ( 3 ) , 1 + 2 ln M max 1 Γ ( α 1 ) , 1 Γ ( n 2 ) , , 1 Γ ( 2 ) = max 3 Γ ( α ) , 3 2 + 2 ln M max 1 Γ ( α 1 ) , 1 = Ξ 2 ,
we conclude that
u U Ξ 2 D 1 + α H u V + G + ( L + G ln M ) Ξ 1 .
On the other hand, because A u = ν B u , by ( I 2 ) , we have
D 1 + α H u V B u V Ξ 0 u U + p n + 1 V .
Hence, according to (56) and (57), we deduce
u U Ξ 2 ( Ξ 0 u U + p n + 1 V ) + G + ( L + G ln M ) Ξ 1 = Ξ 2 Ξ 0 u U + Ξ 2 p n + 1 V + G + ( L + G ln M ) Ξ 1 .
So, by using (58) and (44) (from ( I 8 ) ), we obtain
u U Ξ 2 p n + 1 V + G + ( L + G ln M ) Ξ 1 1 Ξ 0 Ξ 2 ,
which means that Λ 1 is bounded. □
 Lemma 10. 
Assume that ( I 1 ) , ( I 2 ) , ( I 6 ) , and ( I 7 ) hold. Then, the set Λ 2 (defined in Lemma 7) is bounded in U .
Proof. 
For any u Λ 2 , u ( θ ) = a 1 ( ln θ ) α 1 + a 2 ( ln θ ) α 2 with a 1 , a 2 R and C 1 ( B u ) = C 2 ( B u ) = 0 . Using the same arguments as those used in the proof of Lemma 7, we prove that | a 1 | and | a 2 | are bounded and, therefore, Λ 2 is bounded.
According to ( I 6 ) and ( I 7 ) , we deduce that there exist the constants θ 3 [ 1 , ) and θ 4 [ 1 , M ] such that | D 1 + α 1 H u ( θ 3 ) | G and | D 1 + α 2 H u ( θ 4 ) | G , that is,
| D 1 + α 1 H u ( θ 3 ) | = | a 1 | Γ ( α ) G , | D 1 + α 2 H u ( θ 4 ) | = | a 1 Γ ( α ) ln θ 4 + a 2 Γ ( α 1 ) | L .
Therefore, we obtain
| a 1 | G Γ ( α ) , | a 2 | L + | a 1 | Γ ( α ) ln M Γ ( α 1 ) L + G ln M Γ ( α 1 ) .
 Lemma 11. 
Assume that ( I 1 ) , ( I 2 ) , ( I 6 ) , and ( I 7 ) hold. Then, the set
Λ 4 = { u K e r A ; ς ζ H ˜ u + ( 1 ζ ) F B u = 0 , ζ [ 0 , 1 ] } U
 is bounded in U , where
ς = 1 , i f ( 40 ) a n d ( 42 ) h o l d C a s e ( I ) , 1 , i f ( 41 ) a n d ( 43 ) h o l d C a s e ( I I ) , 1 , i f ( 41 ) a n d ( 42 ) h o l d C a s e ( I I I ) , 1 , i f ( 40 ) a n d ( 43 ) h o l d C a s e ( I V ) ,
and H ˜ : K e r A I m F is the linear isomorphism operator defined by
H ˜ ( u ( θ ) ) = 1 Δ ( d b 2 b b 1 ) 1 θ 2 + 1 Δ ( a b 1 c b 2 ) ln θ θ 2 , i n t h e c a s e s ( I ) , ( I I ) , 1 Δ ( d b 2 + b b 1 ) 1 θ 2 + 1 Δ ( a b 1 c b 2 ) ln θ θ 2 , i n t h e c a s e s ( I I I ) , ( I V ) ,
for u K e r A , u ( θ ) = b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 , θ 1 , with b 1 , b 2 R .
Proof. 
We prove this lemma, without loss of generality, in the case of ( I ) —specifically, (40) and (42) hold, and ς = 1 . Let u Λ 4 , u ( θ ) = b 1 ( ln θ ) α 1 + b 2 ( ln θ ) α 2 with b 1 , b 2 R and
ζ H ˜ u ( θ ) = ( 1 ζ ) F B u ( θ ) , ζ [ 0 , 1 ] .
As in Lemma 8, we can show that | b 1 | and | b 2 | are bounded when ζ = 0 or ζ = 1 . Now, we prove that | b 1 | and | b 2 | are bounded in the case of ζ ( 0 , 1 ) . According to relation (59), we find
ζ ( d b 2 b b 1 ) = ( 1 ζ ) ( d C 1 B u b C 2 B u ) , ζ ( a b 1 c b 2 ) = ( 1 ζ ) ( a C 2 B u c C 1 B u ) .
Since Δ 0 , from the above system, we obtain
ζ b 1 = ( 1 ζ ) C 2 B u ,
ζ b 2 = ( 1 ζ ) C 1 B u .
According to (40) and (60), we deduce that | b 1 | Γ ( α ) G , so | b 1 |   G Γ ( α ) . We use the method of reduction to the absurd. If we suppose that | b 1 | Γ ( α ) > G , then | H D 1 + α 1 u ( θ ) | = | b 1 | Γ ( α ) > G . Now, using (40), we obtain s g n { H D 1 + α 1 u ( θ ) } C 2 B u ( θ ) > 0 for all θ [ 1 , ) or s g n { b 1 } C 2 B u ( θ ) > 0 for all θ [ 1 , ) . Then, by using (60), we conclude
0 ζ b 1 s g n { b 1 } = ζ b 1 s g n { H D 1 + α 1 u ( θ ) } = ( 1 ζ ) C 2 B u s g n { H D 1 + α 1 u ( θ ) } < 0 ,
which is a contradiction.
Next, according to (42) and (61), we deduce that | b 2 | Γ ( α 1 ) L + G ln M ; then, | b 2 |   1 Γ ( α 1 ) ( L + G ln M ) . We also use the method of reduction to the absurd. We suppose that | b 2 | Γ ( α 1 ) > L + G ln M . So, we find
| H D 1 + α 2 u ( M ) | = | b 1 Γ ( α ) ln M + b 2 Γ ( α 1 ) | | b 1 | Γ ( α ) ln M + | b 2 | Γ ( α 1 ) > L + G ln M | b 1 | Γ ( α ) ln M L + G ln M G ln M = L .
So, according to (42), we have s g n { H D 1 + α 2 u ( θ ) } C 1 B u ( θ ) > 0 for all θ [ 1 , M ] . For θ = 1 , we obtain s g n { H D 1 + α 2 u ( 1 ) } C 1 B u ( 1 ) > 0 . Hence, by using (61), we deduce
0 ζ b 2 s g n { b 2 } = ζ b 2 s g n { H D 1 + α 2 u ( 1 ) } = ( 1 ζ ) C 1 B u s g n { H D 1 + α 2 u ( 1 ) } < 0 ,
which is a contradiction.
So, we find that set Λ 4 is bounded. □
We now present the proof of Theorem 2.
Proof of Theorem 2. 
Let Λ 5 U be a bounded open set such that Λ ¯ 1 Λ ¯ 2 Λ ¯ 4 Λ 5 . According to Lemmas 3 and 5, operator A is a Fredholm operator of index 0 and operator B is A -compact on Λ ¯ 5 . According to Lemmas 9 and 10, we see that conditions ( i ) and ( i i ) of the Mawhin continuation theorem are satisfied. In what follows, we prove that condition ( i i i ) is also satisfied. For this, we define the following operator:
T ( u , ζ ) = ς ζ H ˜ ( u ) + ( 1 ζ ) F B u ,
where ς is given in Lemma 11. By use of Lemma 11, we have T ( u , ζ ) 0 for all u K e r A Λ 5 and ζ [ 0 , 1 ] . Then, according to the homotopy of degree, we obtain
d e g { F B | K e r A , Λ 5 K e r A , 0 } = d e g { T ( · , 0 ) , Λ 5 K e r A , 0 } = d e g { T ( · , 1 ) , Λ 5 K e r A , 0 } = d e g { ς H ˜ , Λ 5 K e r A , 0 } 0 .
Therefore, by applying Theorem IV.13 from [1], we deduce that the equation A u = B u , that is, Equation (11), has at least one solution in D ( A ) Λ ¯ 5 . This signifies that problem (1), (2) has at least one solution in space U . □

4. Examples

 Example 1. 
Let α = 11 3 , n = 4 , A = B = 20 ,
P ( σ ) = a 0 ( σ 1 ) + 1 6 , σ [ 1 , 10 ) , 7 6 , σ [ 10 , 20 ] , Q ( σ ) = 1 , σ [ 1 , 10 ) , 1 100 ( σ 10 ) 2 + 1 , σ [ 10 , 20 ] ,
where a 0 = ln 10 9 ln 10 .
We consider the following fractional differential equation:
D 1 + 11 / 3 H u ( θ ) = f ( θ , u ( θ ) , D 1 + 2 / 3 H u ( θ ) , D 1 + 5 / 3 H u ( θ ) , D 1 + 8 / 3 H u ( θ ) ) , θ ( 1 , ) ,
subject to the following boundary conditions:
u ( 1 ) = u ( 1 ) = 0 , D 1 + 5 / 3 H u ( θ ) = a 0 1 10 D 1 + 5 / 3 H u ( σ ) d σ + ( 1 9 a 0 ) D 1 + 5 / 3 H u ( 10 ) , D 1 + 8 / 3 H u ( ) = 10 20 1 50 ( σ 10 ) D 1 + 8 / 3 H u ( σ ) d ( σ ) .
After some computations, we find a 0.17788008 , b 0.13237075 , c 0.00193147 , d 0.00628112 , and Δ 0.00137296 0 . In addition, we have 1 20 d P ( σ ) = 1 , 1 20 d Q ( σ ) = 1 , and 1 20 ln σ d P ( σ ) = 0 . So, assumption ( I 1 ) is satisfied.
We consider the following function for x i R , i = 1 , , 4 :
f ( θ , x 1 , x 2 , x 3 , x 4 ) = e 4 θ + 3 θ 2 + 1 2 θ + 1 3 arctan x 1 x 2 4 ( x 2 2 + 1 ) + x 3 30 , θ [ 1 , 10 ] , c 0 x 4 5 θ 2 , θ ( 10 , ) ,
where
c 0 = 25 2 ( 2 ln 2 1 ) 1 10 a 0 ln 10 τ 10 + τ + ln 10 τ e 4 τ + 3 d τ τ 0.4106 .
We obtain
| f ( θ , x 1 , x 2 , x 3 , x 4 ) | p 1 ( θ ) | x 1 | 1 + ( ln θ ) 8 / 3 + p 2 ( θ ) | x 2 | 1 + ( ln θ ) 2 + p 3 ( θ ) | x 3 | 1 + ln θ + p 4 ( θ ) | x 4 | + p 5 ( θ ) , θ [ 1 , ) , x i R , i = 1 , , 4 ,
with
p 1 ( θ ) = 1 3 e 4 θ + 3 ( 1 + ( ln θ ) 8 / 3 ) , θ [ 1 , 10 ] , 0 , θ ( 10 , ) , p 2 ( θ ) = 1 4 e 4 θ + 3 ( 1 + ( ln θ ) 2 ) , θ [ 1 , 10 ] , 0 , θ ( 10 , ) , p 3 ( θ ) = { 1 30 e 4 θ + 3 ( 1 + ln θ ) , θ [ 1 , 10 ] , 0 , θ ( 10 , ) , p 4 ( θ ) = { 0 , θ [ 1 , 10 ] , c 0 5 θ 2 , θ ( 10 , ) , p 5 ( θ ) = { e 4 θ + 3 θ 2 + 1 2 θ , θ [ 1 , 10 ] , 0 , θ ( 10 , ) .
We have p i L 1 ( 1 , ) for i = 1 , , 5 and p 1 V   0.03200099 , p 2 V   0.02468634 , p 3 V   0.00369825 , p 4 V   0.00821168 , and p 5 V   0.09543632 . We also find Ξ 0 0.06859726 , and Ξ 1 = 1 . So, assumption ( I 2 ) is also satisfied.
We now verify assumption ( I 3 ) . We consider E = 200 and F = 10 . Let u D ( A ) K e r A .
If | H D 1 + 5 / 3 u ( θ ) | > E , then we obtain
Θ 1 = C 1 ( B u ) = 1 20 1 σ ln σ τ B u ( τ ) τ d τ d P ( σ ) = a 0 1 10 1 σ ln σ τ B u ( τ ) τ d τ d σ + ( 1 9 a 0 ) 1 10 ln 10 τ B u ( τ ) τ d τ = a 0 1 10 [ 1 σ ln σ τ 1 τ e 4 τ + 3 τ 2 + 1 2 τ + 1 3 arctan u ( τ ) D 1 + 2 / 3 H u ( τ ) 4 ( ( D 1 + 2 / 3 H u ( τ ) ) 2 + 1 ) + 1 30 D 1 + 5 / 3 H u ( τ ) d τ d σ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 τ 2 + 1 2 τ + 1 3 arctan u ( τ ) D 1 + 2 / 3 H u ( τ ) 4 ( ( D 1 + 2 / 3 H u ( τ ) ) 2 + 1 ) + 1 30 D 1 + 5 / 3 H u ( τ ) d τ = a 0 1 10 1 σ ln σ τ 1 τ e 4 τ + 3 h ( τ ) d τ d σ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ + a 0 1 10 1 σ ln σ τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ d σ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ ,
where
h ( τ ) = τ 2 + 1 2 τ + 1 3 arctan u ( τ ) D 1 + 2 / 3 H u ( τ ) 4 ( ( D 1 + 2 / 3 H u ( τ ) ) 2 + 1 ) , τ [ 1 , 10 ] .
Therefore, we find
Θ 1 = C 1 ( B u ) = a 0 1 10 τ 10 ln σ τ 1 τ e 4 τ + 3 h ( τ ) d σ d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ + a 0 1 10 τ 10 ln σ τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d σ d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ = a 0 1 10 1 τ e 4 τ + 3 h ( τ ) τ 10 ln σ τ d σ d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ + a 0 1 10 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) τ 10 ln σ τ d σ d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ
= a 0 1 10 1 τ e 4 τ + 3 10 ln 10 τ 10 + τ h ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ + a 0 1 10 1 30 τ e 4 τ + 3 10 ln 10 τ 10 + τ D 1 + 5 / 3 H u ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ = 1 10 V ( τ ) h ( τ ) d τ + 1 30 1 10 V ( τ ) D 1 + 5 / 3 H u ( τ ) d τ ,
where
V ( τ ) = a 0 τ e 4 τ + 3 10 ln 10 τ 10 + τ + ( 1 9 a 0 ) ln 10 τ 1 τ e 4 τ + 3 , τ [ 1 , 10 ] .
We see that V ( τ ) 0 for all τ [ 1 , 10 ] and h ( τ ) [ h 1 ( τ ) , h 2 ( τ ) ] , where h 1 ( τ ) = τ 2 + 1 2 τ π 6 1 8 and h 2 ( τ ) = τ 2 + 1 2 τ + π 6 + 1 8 for τ [ 1 , 10 ] . We have min σ [ 1 , 10 ] h 1 ( σ ) 0.3514 > 0 and V ( τ ) h 2 ( τ ) V ( τ ) h ( τ ) V ( τ ) h 1 ( τ ) for all τ [ 1 , 10 ] .
If D 1 + 5 / 3 H u ( θ ) > E , then Θ 1 < E 1 10 1 30 V ( τ ) d τ + 1 10 V ( τ ) h 1 ( τ ) d τ 0.0898 < 0 .
If D 1 + 5 / 3 H u ( θ ) < E , then Θ 1 > E 1 10 1 30 V ( τ ) d τ + 1 10 V ( τ ) h 2 ( τ ) d τ 0.06283 > 0 .
So, Θ 1 0 .
If | D 1 + 8 / 3 H u ( θ ) | > E , we obtain
Θ 2 = C 2 ( B u ) = 1 20 σ B u ( τ ) τ d τ d Q ( σ ) = 10 20 1 50 ( σ 10 ) σ B u ( τ ) τ d τ d σ = 10 20 1 50 ( σ 10 ) σ c 0 5 τ 3 D 1 + 8 / 3 H u ( τ ) d τ d σ .
If D 1 + 8 / 3 H u ( θ ) > E , then Θ 2 > E 10 20 1 50 ( σ 1 ) σ c 0 5 τ 3 d τ d σ > 0 .
If D 1 + 8 / 3 H u ( θ ) < E , then Θ 2 < E 10 20 1 50 ( σ 1 ) σ c 0 5 τ 3 d τ d σ < 0 .
So, Θ 2 0 .
We deduce that assumption ( I 3 ) is satisfied.
In what follows, we verify assumption ( I 4 ) . We consider r = 50 and u ( θ ) = b 1 ( ln θ ) 8 / 3 + b 2 ( ln θ ) 5 / 3 for θ [ 1 , ) , with b 1 , b 2 R and | b 1 | > r or | b 2 | > r . We obtain
D 1 + 8 / 3 H u ( θ ) = b 1 Γ 11 3 , D 1 + 5 / 3 H u ( θ ) = b 1 Γ 11 3 ln θ + b 2 Γ 8 3 , D 1 + 2 / 3 H u ( θ ) = b 1 2 Γ 11 3 ( ln θ ) 2 + b 2 Γ 8 3 ln θ .
Then, we find
Θ 0 = b 1 C 1 ( B u ) ( θ ) + b 2 C 2 ( B u ) ( θ ) = b 1 1 20 1 σ ln σ τ B u ( τ ) τ d τ d P ( σ ) + b 2 1 20 σ B u ( τ ) τ d τ d Q ( σ ) = b 1 a 0 1 10 1 σ ln σ τ B u ( τ ) τ d τ d σ + ( 1 9 a 0 ) 1 10 ln 10 τ B u ( τ ) τ d τ + b 2 10 20 1 50 ( σ 10 ) σ B u ( τ ) τ d τ d σ = b 1 a 0 1 10 1 σ ln σ τ 1 τ e 4 τ + 3 h ( τ ) + 1 30 D 1 + 5 / 3 H u ( τ ) d τ d σ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) + 1 30 D 1 + 5 / 3 H u ( τ ) d τ + b 2 10 20 1 50 ( σ 10 ) σ c 0 5 τ 3 D 1 + 8 / 3 H u ( τ ) d τ d σ = b 1 a 0 1 10 1 σ ln σ τ 1 τ e 4 τ + 3 h ( τ ) d τ d σ + a 0 1 10 1 σ ln σ τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ d σ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ + b 2 10 20 1 50 ( σ 10 ) σ c 0 5 τ 3 D 1 + 8 / 3 H u ( τ ) d τ d σ .
So, we obtain
Θ 0 = b 1 a 0 1 10 τ 10 ln σ τ d σ 1 τ e 4 τ + 3 h ( τ ) d τ + a 0 1 10 τ ln σ τ d σ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 D 1 + 5 / 3 H u ( τ ) d τ + b 2 10 20 1 50 ( σ 10 ) σ c 0 5 τ 3 D 1 + 8 / 3 H u ( τ ) d τ d σ = b 1 a 0 1 10 10 ln 10 τ 10 + τ 1 τ e 4 τ + 3 h ( τ ) d τ + a 0 1 10 10 ln 10 τ 10 + τ 1 30 τ e 4 τ + 3 b 1 Γ 11 3 ln τ + b 2 Γ 8 3 d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 b 1 Γ 11 3 ln τ + b 2 Γ 8 3 d τ + b 2 10 20 1 50 ( σ 10 ) σ c 0 5 τ 3 b 1 Γ 11 3 d τ d σ = C 1 b 1 2 + C 2 b 1 = b 1 ( C 1 b 1 + C 2 ) ,
where
C 1 = a 0 1 10 10 ln 10 τ 10 + τ 1 30 τ e 4 τ + 3 Γ 11 3 ln τ d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 30 τ e 4 τ + 3 Γ 11 3 ln τ d τ , C 2 = a 0 1 10 10 ln 10 τ 10 + τ 1 τ e 4 τ + 3 h ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h ( τ ) d τ .
We obtain C 1 0.00052405 < 0 and C 2 [ C 2 , C 2 ] , where
C 2 = a 0 1 10 10 ln 10 τ 10 + τ 1 τ e 4 τ + 3 h 2 ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h 2 ( τ ) d τ 0.02176398 , C 2 = a 0 1 10 10 ln 10 τ 10 + τ 1 τ e 4 τ + 3 h 1 ( τ ) d τ + ( 1 9 a 0 ) 1 10 ln 10 τ 1 τ e 4 τ + 3 h 1 ( τ ) d τ 0.00530448 .
So, we find C 2 C 1 [ C 2 C 1 , C 2 C 1 ] , with C 2 C 1 10.12211324 and C 2 C 1 41.53048712 . Then, if | b 1 | > 50 , we deduce that Θ 0 < 0 . So, assumption ( I 4 ) is satisfied.
In addition, Ξ 0 ( 2 + Ξ 1 ln F ) 0.29514557 < 1 , that is, ( I 5 ) is also satisfied. Therefore, according to Theorem 1, we conclude that problem (62), (63), with the function f given by (64), has at least one solution.
 Example 2. 
Let α = 9 4 , n = 3 , A = B = 10 , and
P ( σ ) = σ 2 + b 0 ( σ 1 ) , σ [ 1 , 5 ] , 2 , σ [ 5 , 10 ] , Q ( σ ) = 1 4 , σ [ 1 , 5 ) , 2 15 ( σ 10 ) + 5 4 , σ [ 5 , 10 ] ,
where b 0 = 2 ln 5 12 4 ln 5 3.6732 .
We consider the following fractional differential equation:
D 1 + 9 / 4 H u ( θ ) = f θ , u ( θ ) , D 1 + 1 / 4 H u ( θ ) , D 1 + 5 / 4 H u ( θ ) , θ ( 1 , ) ,
supplemented with the following boundary conditions:
u ( 1 ) = 0 , D 1 + 1 / 4 H u ( 1 ) = 1 5 ( 2 σ + b 0 ) D 1 + 1 / 4 H u ( σ ) d σ ( 23 + 4 b 0 ) D 1 + 1 / 4 H u ( 5 ) , D 1 + 5 / 4 H u ( ) = 1 3 D 1 + 5 / 4 H u ( 5 ) + 2 15 5 10 D 1 + 5 / 4 H u ( σ ) d σ .
After some computations, we obtain a 0.26300051 , b 0.18818034 , c 0.01333333 , d 0.03017152 , and Δ 0.01044419 0 . We also find 1 10 d P ( σ ) = 1 , 1 10 d Q ( σ ) = 1 , and 1 10 ln σ d P ( σ ) = 0 . So, assumption ( I 1 ) is satisfied.
We consider the following function:
f ( θ , x 1 , x 2 , x 3 ) = 1 10 e 2 θ + 1 sin x 1 1 + ( ln θ ) 5 / 4 + 1 20 g ( t ) e 3 θ + 2 x 2 + 1 17 ( 1 g ( θ ) ) e θ + 3 x 3 + 1 12 θ 2 , θ [ 1 , ) , x i R , i = 1 , , 3 ,
where
g ( θ ) = 1 , θ [ 1 , 5 ] , 0 , θ ( 5 , ) .
We obtain
| f ( θ , x 1 , x 2 , x 3 ) | p 1 ( θ ) | x 1 | 1 + ( ln θ ) 5 / 4 + p 2 ( θ ) | x 2 | 1 + ln θ + p 3 ( θ ) | x 3 | + p 4 ( θ ) ,
for all θ [ 1 , ) and x i R , i = 1 , , 3 , with
p 1 ( θ ) = 1 10 e 2 θ + 1 , p 2 ( θ ) = 1 20 e 3 θ + 2 ( 1 + ln θ ) , θ [ 1 , 5 ] , 0 , θ ( 5 , ) , p 3 ( θ ) = 0 , θ [ 1 , 5 ] , 1 17 e θ + 3 , θ ( 5 , ) , p 4 ( θ ) = 1 12 θ 2 .
We have p i L 1 ( 1 , ) for i = 1 , , 4 and p 1 V 0.01839397 , p 2 V 0.00773814 , and p 3 V 0.00796089 . In addition, we find Ξ 0 0.03409301 , Ξ 1 1.10326265 , Ξ 2 6.19909584 , and Ξ 0 Ξ 2 0.21134587 < 1 . So, assumptions ( I 2 ) and ( I 8 ) are satisfied.
We now verify assumption ( I 7 ) . We consider L = 25 and M = 5 . Let u D ( A ) satisfy | H D 1 + 1 / 4 u ( θ ) | > 25 for all θ [ 1 , 5 ] . Then, we obtain
C 1 B u ( θ ) = 1 10 1 σ ln σ τ B u ( τ ) τ d τ d P ( σ ) = 1 5 ( 2 σ + b 0 ) 1 σ ln σ τ B u ( τ ) τ d τ d σ ( 23 + 4 b 0 ) 1 5 ln 5 τ B u ( τ ) τ d τ = 1 5 τ 5 ( 2 σ + b 0 ) ln σ τ d σ B u ( τ ) τ d τ ( 23 + 4 b 0 ) 1 5 ln 5 τ B u ( τ ) τ d τ = 1 5 τ 5 ( 2 σ + b 0 ) ln σ τ d σ 1 τ 1 10 e 2 τ + 1 sin u ( τ ) 1 + ( ln τ ) 5 / 4 + 1 12 τ 2 + 1 20 e 3 τ + 2 D 1 + 1 / 4 H u ( τ ) d τ ( 23 + 4 b 0 ) 1 5 ln 5 τ 1 τ 1 10 e 2 τ + 1 sin u ( τ ) 1 + ( ln τ ) 5 / 4 + 1 12 τ 2 + 1 20 e 3 τ + 2 D 1 + 1 / 4 H u ( τ ) d τ = 1 5 τ 5 ( 2 σ + b 0 ) ln σ τ d σ 1 τ k ( τ ) d τ + 1 5 τ 5 ( 2 σ + b 0 ) ln σ τ d σ 1 20 τ e 3 τ + 2 D 1 + 1 / 4 H u ( τ ) d τ ( 23 + 4 b 0 ) 1 5 ln 5 τ 1 τ k ( τ ) d τ ( 23 + 4 b 0 ) 1 5 ln 5 τ 1 20 τ e 3 τ + 2 D 1 + 1 / 4 H u ( τ ) d τ ,
where
k ( τ ) = 1 10 e 2 τ + 1 sin u ( τ ) 1 + ( ln τ ) 5 / 4 + 1 12 τ 2 , τ [ 1 , 5 ] .
Then, we find
C 1 B u ( θ ) = 1 5 U ( τ ) ( 23 + 4 b 0 ) ln 5 τ 1 τ k ( τ ) d τ + 1 5 U ( τ ) ( 23 + 4 b 0 ) ln 5 τ 1 20 τ e 3 τ + 2 D 1 + 1 / 4 H u ( τ ) d τ ,
where U ( τ ) = ( 25 + 5 b 0 ) ln 5 τ 25 2 5 b 0 + τ 2 2 + b 0 τ for τ [ 1 , 5 ] .
We have k 1 ( τ ) k ( τ ) k 2 ( τ ) for all τ [ 1 , 5 ] , where k 1 ( τ ) = 1 10 e 2 τ + 1 + 1 12 τ 2 and k 2 ( τ ) = 1 10 e 2 τ + 1 + 1 12 τ 2 for all τ [ 1 , 5 ] . In addition, we find U ( τ ) 0 for all τ [ 1 , 5 ] and inf τ [ 1 , 5 ] k 1 ( τ ) 0.0033 > 0 , so k 1 ( τ ) > 0 for all τ [ 1 , 5 ] . Then we have the following inequalities:
k 2 ( τ ) U ( τ ) k ( τ ) U ( τ ) k 1 ( τ ) U ( τ ) , τ [ 1 , 5 ] .
If D 1 + 1 / 4 H u ( θ ) > L , then
C 1 B u ( θ ) 1 5 U ( τ ) 1 τ k 1 ( τ ) d τ + L 1 5 U ( τ ) 1 20 τ e 3 τ + 2 d τ < 0 , θ [ 1 , 5 ] ,
so s g n { D 1 + 1 / 4 H u ( θ ) } C 1 B u ( θ ) < 0 for all θ [ 1 , 5 ] .
If D 1 + 1 / 4 H u ( θ ) < L , then
C 1 B u ( θ ) 1 5 U ( τ ) 1 τ k 2 ( τ ) d τ L 1 5 U ( τ ) 1 20 τ e 3 τ + 2 d τ 0.02701926 + 0.00135717 L > 0 ,
therefore, s g n { D 1 + 1 / 4 H u ( θ ) } C 1 B u ( θ ) < 0 for all θ [ 1 , 5 ] .
We deduce that assumption ( I 7 ) is satisfied.
Finally, we verify assumption ( I 6 ) . We consider G = 5 . Let u D ( A ) , which satisfies | D 1 + 5 / 4 H u ( θ ) | > G for all θ [ 1 , ) . Then, we obtain
C 2 B u ( θ ) = 1 10 σ B u ( τ ) τ d τ d Q ( σ ) = 1 3 5 B u ( τ ) τ d τ + 2 15 5 10 σ B u ( τ ) τ d τ d σ = 1 3 5 B u ( τ ) τ d τ + 2 15 5 10 σ 10 B u ( τ ) τ d τ d σ + 5 10 10 B u ( τ ) τ d τ d σ = 1 3 5 B u ( τ ) τ d τ + 2 15 5 10 5 τ d σ B u ( τ ) τ d τ + 10 B u ( τ ) τ d τ 5 10 d σ = 1 3 5 B u ( τ ) τ d τ + 2 15 5 10 ( τ 5 ) B u ( τ ) τ d τ + 5 10 B u ( τ ) τ d τ = 1 3 5 B u ( τ ) τ d τ + 2 15 5 10 ( τ 5 ) B u ( τ ) τ d τ + 2 3 10 B u ( τ ) τ d τ .
So, we find
C 2 B u ( θ ) = 1 3 5 1 τ k ( τ ) + 1 17 e τ + 3 D 1 + 5 / 4 H u ( τ ) d τ + 2 15 5 10 τ 5 τ k ( τ ) + 1 17 e τ + 3 D 1 + 5 / 4 H u ( τ ) d τ + 2 3 10 1 τ k ( τ ) + 1 17 e τ + 3 D 1 + 5 / 4 H u ( τ ) d τ = 1 3 5 1 τ k ( τ ) d τ + 2 15 5 10 τ 5 τ k ( τ ) d τ + 2 3 10 1 τ k ( τ ) d τ + 1 3 5 1 17 τ e τ + 3 D 1 + 5 / 4 H u ( τ ) d τ + 2 255 5 10 τ 5 τ e τ + 3 D 1 + 5 / 4 H u ( τ ) d τ + 2 3 10 1 17 τ e τ + 3 D 1 + 5 / 4 H u ( τ ) d τ .
If D 1 + 5 / 4 H u ( θ ) > G , then we obtain
C 2 B u ( θ ) 1 3 5 1 τ k 1 ( τ ) d τ + 2 15 5 10 τ 5 τ k 1 ( τ ) d τ + 2 3 10 1 τ k 1 ( τ ) d τ + G 1 3 5 1 17 τ e τ + 3 d τ + 2 255 5 10 τ 5 τ e τ + 3 d τ + 2 3 10 1 17 τ e τ + 3 d τ > 0 ,
so s g n { D 1 + 5 / 4 H u ( θ ) } C 2 B u ( θ ) > 0 for all θ [ 1 , ) .
If D 1 + 5 / 4 H u ( θ ) < G , then we find
C 2 B u ( θ ) 1 3 5 1 τ k 2 ( τ ) d τ + 2 15 5 10 τ 5 τ k 2 ( τ ) d τ + 2 3 10 1 τ k 2 ( τ ) d τ G 1 3 5 1 17 τ e τ + 3 d τ + 2 255 5 10 τ 5 τ e τ + 3 d τ + 2 3 10 1 17 τ e τ + 3 d τ 0.00111156 0.00060861 G < 0 ,
so s g n { D 1 + 5 / 4 H u ( θ ) } C 2 B u ( θ ) > 0 for all θ [ 1 , ) .
We conclude that assumption ( I 6 ) is satisfied.
Therefore, according to Theorem 2, we deduce that the boundary value problem (65), (66), with the function f given by (67), has at least one solution.

5. Conclusions

In this paper, we studied the Hadamard fractional differential Equation (1) defined on the infinite interval of ( 1 , ) , which involves fractional derivatives of various orders, supplemented with the integral boundary conditions (2). These nonlocal conditions encompass Hadamard fractional derivatives and Riemann–Stieltjes integrals. Problem (1), (2) is a resonant problem, because the corresponding homogeneous boundary value problem admits nontrivial solutions. We first expressed our problem as an operator equation in suitable Banach spaces containing two operators whose key properties were analyzed in detail. Subsequently, we established two existence theorems for the solutions of (1), (2) by employing the coincidence degree theory developed by J. Mawhin—more concretely, the Mawhin continuation theorem.

Author Contributions

Conceptualization, R.L.; Formal analysis, A.T. and R.L.; Methodology, A.T. and R.L.; Writing–original draft preparation, A.T. and R.L.; Writing–review and editing, A.T. and R.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Acknowledgments

The authors thank the referees for their valuable comments and suggestions.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Mawhin, J. Topological Degree Methods in Nonlinear Boundary Value Problems; Regional Conference Series in Mathematics, No. 40; American Mathematical Society: Providence, RI, USA, 1979. [Google Scholar]
  2. Mawhin, J. Topological Degree and Boundary Value Problems for Nonlinear Differential Equations. In Topological Methods for Ordinary Differential Equations; Lecture Notes in Mathematics; Springer: Berlin/Heidelberg, Germany, 1993; Volume 1537, pp. 74–142. [Google Scholar]
  3. Luca, R.; Tudorache, A. Existence of solutions for a Hadamard fractional boundary value problem at resonance. Fractal Fract. 2025, 9, 119. [Google Scholar] [CrossRef]
  4. Bohner, M.; Domoshnitsky, A.; Padhi, S.; Srivastava, S.N. Existence of solutions by coincidence degree theory for Hadamard fractional differential equations at resonance. Turk. J. Math. 2024, 48, 296–306. [Google Scholar] [CrossRef]
  5. Zhang, W.; Liu, W. Existence of solutions for fractional multi-point boundary value problems on an infinite interval at resonance. Mathematics 2020, 8, 126. [Google Scholar] [CrossRef]
  6. Djebali, S.; Aoun, A.G. Resonant fractional differential equations with multi-point boundary conditions on (0,+). J. Nonlinear Funct. Anal. 2019, 2019, 21. [Google Scholar]
  7. Imaga, O.F.; Iyase, S.A. On a fractional-order p-Laplacian boundary value problem at resonance on the half-line with two dimensional kernel. Adv. Differ. Equ. 2021, 2021, 13–14. [Google Scholar] [CrossRef]
  8. Zhang, W.; Liu, W.; Chen, T. Solvability for a fractional p-Laplacian multipoint boundary value problem at resonance on infinite interval. Adv. Differ. Equ. 2016, 2016, 183. [Google Scholar] [CrossRef]
  9. Ge, F.D.; Zhou, H.C.; Kou, C.H. Existence of solutions for a coupled fractional differential equations with infinitely many points boundary conditions at resonance on an unbounded domain. Differ. Equ. Dyn. Syst. 2019, 27, 395–411. [Google Scholar] [CrossRef]
  10. Domoshnitsky, A.; Srivastava, S.N.; Padhi, S. Existence of solutions for a higher order Riemann-Liouville fractional differential equation by Mawhin’s coincidence degree theory. Math. Meth. Appl. Sci. 2023, 46, 12018–12034. [Google Scholar] [CrossRef]
  11. Wang, Y. Necessary conditions for the existence of positive solutions to fractional boundary value problems at resonance. Appl. Math. Lett. 2019, 97, 34–40. [Google Scholar] [CrossRef]
  12. Wang, Y.; Huang, Y.; Li, X. Positive solutions for fractional differential equation at resonance under integral boundary conditions. Demonstr. Math. 2022, 55, 238–253. [Google Scholar] [CrossRef]
  13. Wang, Y.; Wu, Y. Positive solutions of fractional differential equation boundary value problems at resonance. J. Appl. Anal. Comput. 2020, 10, 2459–2475. [Google Scholar]
  14. Baitiche, Z.; Guerbati, K.; Hammouche, H.; Benchohra, M.; Graef, J. Sequential fractional differential equations at resonance. Funct. Differ. Equ. 2020, 26, 167–184. [Google Scholar] [CrossRef]
  15. Cheng, L.; Liu, W.; Ye, Q. Boundary value problem for a coupled system of fractional differential equations with p-Laplacian operator at resonance. Electr. J. Differ. Equ. 2014, 2014, 1–12. [Google Scholar]
  16. Hu, L. Existence of solutions to a coupled system of fractional differential equations with infinite-point boundary value conditions at resonance. Adv. Differ. Equ. 2016, 2016, 200. [Google Scholar] [CrossRef]
  17. Ma, W.; Meng, S.; Cui, Y. Resonant integral boundary value problems for Caputo fractional differential equations. Math. Probl. Eng. 2018, 2018, 5438592. [Google Scholar] [CrossRef]
  18. Garra, R.; Mainardi, F.; Spada, G. A generalization of the Lomnitz logarithmic creep law via Hadamard fractional calculus. Chaos Solitons Fractals 2017, 102, 333–338. [Google Scholar] [CrossRef]
  19. Garra, R.; Orsingher, E.; Polito, F. A note on Hadamard fractional differential equations with varying coefficients and their applications in probability. Mathematics 2018, 6, 4. [Google Scholar] [CrossRef]
  20. Ahmad, B.; Alsaedi, A.; Ntouyas, S.K.; Tariboon, J. Hadamard-Type Fractional Differential Equations; Inclusions and Inequalities; Springer: Cham, Switzerland, 2017. [Google Scholar]
  21. Luca, R.; Tudorache, A. On a system of Hadamard fractional differential equations with nonlocal boundary conditions on an infinite interval. Fractal Fract. 2023, 7, 458. [Google Scholar] [CrossRef]
  22. Tudorache, A.; Luca, R. Positive solutions for a system of Hadamard fractional boundary value problems on an infinite interval. Axioms 2023, 12, 793. [Google Scholar] [CrossRef]
  23. Tudorache, A.; Luca, R. Positive solutions to a system of coupled Hadamard fractional boundary value problems. Fractal Fract. 2024, 8, 543. [Google Scholar] [CrossRef]
  24. Agarwal, R.P.; O’Regan, D. Infinite Interval Problems for Differential, Difference and Integral Equations; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2001. [Google Scholar]
  25. Kilbas, A.A.; Srivastava, H.M.; Trujillo, J.J. Theory and Applications of Fractional Differential Equations, North-Holland Mathematics Studies; Elsevier Science B.V.: Amsterdam, The Netherlands, 2006; Volume 204. [Google Scholar]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Tudorache, A.; Luca, R. A Hadamard Fractional Boundary Value Problem on an Infinite Interval at Resonance. Fractal Fract. 2025, 9, 378. https://doi.org/10.3390/fractalfract9060378

AMA Style

Tudorache A, Luca R. A Hadamard Fractional Boundary Value Problem on an Infinite Interval at Resonance. Fractal and Fractional. 2025; 9(6):378. https://doi.org/10.3390/fractalfract9060378

Chicago/Turabian Style

Tudorache, Alexandru, and Rodica Luca. 2025. "A Hadamard Fractional Boundary Value Problem on an Infinite Interval at Resonance" Fractal and Fractional 9, no. 6: 378. https://doi.org/10.3390/fractalfract9060378

APA Style

Tudorache, A., & Luca, R. (2025). A Hadamard Fractional Boundary Value Problem on an Infinite Interval at Resonance. Fractal and Fractional, 9(6), 378. https://doi.org/10.3390/fractalfract9060378

Article Metrics

Back to TopTop