Open Access

The existence of solutions for a nonlinear mixed problem of singular fractional differential equations

Advances in Difference Equations20132013:359

DOI: 10.1186/1687-1847-2013-359

Received: 27 May 2013

Accepted: 28 August 2013

Published: 11 December 2013

Abstract

By using fixed point results on cones, we study the existence of solutions for the singular nonlinear fractional boundary value problem

D α c u ( t ) = f ( t , u ( t ) , u ( t ) , c D β u ( t ) ) , u ( 0 ) = a u ( 1 ) , u ( 0 ) = b c D β u ( 1 ) , u ( 0 ) = u ( 0 ) = u ( n 1 ) ( 0 ) = 0 ,

where n 3 is an integer, α ( n 1 , n ) , 0 < β < 1 , 0 < a < 1 , 0 < b < Γ ( 2 β ) , f is an L q -Caratheodory function, q > 1 α 1 and f ( t , x , y , z ) may be singular at value 0 in one dimension of its space variables x, y, z. Here, c D stands for the Caputo fractional derivative.

Keywords

boundary value problem fixed point fractional differential equation Green function regularization singular

1 Introduction

Fractional differential equations (see, for example, [16] and references therein) started to play an important role in several branches of science and engineering. There are some works about existence of solutions for the nonlinear mixed problems of singular fractional boundary value problem (see, for example, [711] and [12]). Also, there are different methods for solving distinct fractional differential equations (see, for example, [1318] and [19]). By using fixed point results on cones, we focus on the existence of positive solutions for a nonlinear mixed problem of singular fractional boundary value problem. For the convenience of the reader, we present some necessary definitions from fractional calculus theory (see, for example, [20]). The Caputo derivative of fractional order α for a function f : [ 0 , ) R is defined by
D α c f ( t ) = 1 Γ ( n α ) 0 t ( t s ) n α 1 f ( n ) ( s ) d s ( n 1 < α < n , n = [ α ] + 1 ) .
Let q 1 . As you know, L q [ 0 , 1 ] denotes the space of functions, whose q th powers of modulus are integrable on [ 0 , 1 ] , equipped with the norm x q = ( 0 1 | x ( t ) | q d t ) 1 q . We consider the sup norm
x = sup { | x ( t ) | : t [ 0 , 1 ] }
on the space C [ 0 , 1 ] . Also, AC [ 0 , 1 ] is the set of absolutely continuous functions on [ 0 , 1 ] . Let B be a subset of R 3 . A function f : [ 0 , 1 ] × B R is called an L q -Caratheodory function whenever the real-valued function f ( , x , y , z ) on [ 0 , 1 ] is measurable for all ( x , y , z ) B , the function f ( t , , , ) : B R is continuous for almost all t ( 0 , 1 ] , and for each compact set U B , there exists a function φ u L q [ 0 , 1 ] such that | f ( t , x , y , z ) | φ u ( t ) for almost all t [ 0 , 1 ] and ( x , y , z ) U . Consider the nonlinear fractional boundary value problem
D α u ( t ) = f ( t , u ( t ) , u ( t ) , c D β u ( t ) ) , u ( 0 ) = a u ( 1 ) , u ( 0 ) = b c D β u ( 1 ) , u ( 0 ) = u ( 0 ) = u ( n - 1 ) ( 0 ) = 0 ,
(*)
where n 3 is an integer, α ( n 1 , n ) , 0 < β < 1 , 0 < a < 1 , 0 < b < Γ ( 2 β ) and q > 1 α 1 . We say that the function u : [ 0 , 1 ] R is a positive solution for the problem whenever u > 0 on [ 0 , 1 ] , D α c u is a function in L q [ 0 , 1 ] , and u satisfies the boundary conditions almost everywhere on [ 0 , 1 ] . In this paper, we suppose that f is an L q -Caratheodory function on [ 0 , 1 ] × B , where B = ( 0 , ) × ( 0 , ) × ( 0 , ) , there exists a positive constant m such that m f ( t , x , y , z ) for almost all t [ 0 , 1 ] and ( x , y , z ) B , f satisfies the estimate
f ( t , x , y , z ) h ( x ) + r ( | y | ) + k ( | z | ) + γ ( t ) w ( x , | y | , | z | ) ,
where h , r , k C ( 0 , ) are positive and non-increasing, γ L q [ 0 , 1 ] and w C ( [ 0 , ) × [ 0 , ) × [ 0 , ) ) are positive, w is non-decreasing in all its variables, 0 1 h q ( s α ) d s < , 0 1 r q ( s α 1 ) d s < , 0 1 k q ( s α β ) d s < , and lim x w ( x , x , x ) x = 0 . Since we suppose that problem () is singular, that is, f ( t , x , y , z ) may be singular at the value 0 of its space variables x, y, z, we use regularization and sequential techniques for the existence of positive solutions of the problem. In this way, for each natural number n define the function f n by
f n ( t , x , y , z ) = f ( t , χ n + ( x ) , χ n + ( y ) , χ n + ( z ) )
for all t [ 0 , 1 ] and ( x , y , z ) R 3 , where
χ n + ( u ) = { u , u 1 n , 1 n , u < 1 n .
It is easy to see that each f n is an L q -Caratheodory function on [ 0 , 1 ] × R 3 , m f n ( t , x , y , z ) ,
f n ( t , x , y , z ) h ( 1 n ) + r ( 1 n ) + k ( 1 n ) + γ ( t ) w ( 1 + x , 1 + | y | , 1 + | z | )
and
f n ( t , x , y , z ) h ( x ) + r ( | y | ) + k ( | z | ) + γ ( t ) w ( 1 + x , 1 + | y | , 1 + | z | )

for almost all t [ 0 , 1 ] and all ( x , y , z ) B . In 2012, Agarwal et al. proved the following result.

Lemma 1.1 [7]

Let ρ L q [ 0 , 1 ] and 0 t 1 < t 2 1 . Then we have | 0 t ( t s ) α 2 ρ ( s ) d s | ( t d d ) 1 / p ρ q for all t [ 0 , 1 ] and
| 0 t 2 ( t 2 s ) α 2 ρ ( s ) d s 0 t 1 ( t 1 s ) α 2 ρ ( s ) d s | ( t 1 d + ( t 2 t 1 ) d t 2 d d ) 1 / p ρ q + ( ( t 2 t 1 ) d d ) 1 / p ρ q ,

where d = ( α 2 ) p + 1 .

2 Main results

Now, we are ready to investigate the problem in regular and singular cases. First, we give the following result.

Lemma 2.1 Let y C [ 0 , 1 ] . Then the boundary value problem
D α c u ( t ) = y ( t ) ( t ( 0 , 1 ) ) , u ( 0 ) = a u ( 1 ) , u ( 0 ) = b c D β u ( 1 ) , u ( 0 ) = u ( 0 ) = u ( n 1 ) ( 0 ) = 0
is equivalent to the fractional integral equation u ( t ) = 0 1 G ( t , s ) y ( s ) d s , where
G ( t , s ) = ( t s ) α 1 Γ ( α ) + a Γ ( α β ) ( Γ ( 2 β ) b ) ( 1 s ) α 1 + b Γ ( α ) Γ ( 2 β ) ( a + t a t ) ( 1 s ) α β 1 ( 1 a ) Γ ( α ) Γ ( α β ) ( Γ ( 2 β ) b )
whenever 0 s t 1 and
G ( t , s ) = a Γ ( α β ) ( Γ ( 2 β ) b ) ( 1 s ) α 1 + b Γ ( α ) Γ ( 2 β ) ( a + t a t ) ( 1 s ) α β 1 ( 1 a ) Γ ( α ) Γ ( α β ) ( Γ ( 2 β ) b )

whenever 0 t s 1 .

Proof From D α c u ( t ) = y ( t ) and the boundary conditions, we obtain
u ( t ) = I α y ( t ) + u ( 0 ) + u ( 0 ) t + u ( 0 ) 2 ! t 2 + + u ( n 1 ) ( 0 ) ( n 1 ) ! t n 1 = 1 Γ ( α ) 0 t ( t s ) α 1 y ( s ) d s + u ( 0 ) + u ( 0 ) t .
By properties of the Caputo derivative, we get
D β c u ( t ) = I α β y ( t ) + c D β ( u ( 0 ) + u ( 0 ) t ) = 1 Γ ( α β ) 0 t ( t s ) α β 1 y ( s ) d s + u ( 0 ) t 1 β Γ ( 2 β ) .
Thus, u ( 1 ) = 1 Γ ( α ) 0 1 ( 1 s ) α 1 y ( s ) d s + u ( 0 ) + u ( 0 ) and
D β c u ( 1 ) = 1 Γ ( α β ) 0 1 ( 1 s ) α β 1 y ( s ) d s + u ( 0 ) Γ ( 2 β ) .
By using the boundary conditions u ( 0 ) = a u ( 1 ) and u ( 0 ) = b c D β u ( 1 ) , we get u ( 0 ) = a ( 1 Γ ( α ) 0 1 ( 1 s ) α 1 y ( s ) d s + u ( 0 ) + u ( 0 ) ) and
u ( 0 ) = b ( 1 Γ ( α β ) 0 1 ( 1 s ) α β 1 y ( s ) d s + u ( 0 ) Γ ( 2 β ) ) .
Hence, u ( 0 ) = b Γ ( 2 β ) Γ ( α β ) ( Γ ( 2 β ) b ) 0 1 ( 1 s ) α β 1 y ( s ) d s and
u ( 0 ) = a ( 1 a ) Γ ( α ) 0 1 ( 1 s ) α 1 y ( s ) d s + a b Γ ( 2 β ) ( 1 a ) Γ ( α β ) ( Γ ( 2 β ) b ) 0 1 ( 1 s ) α β 1 y ( s ) d s .
Thus,
u ( t ) = 1 Γ ( α ) 0 t ( t s ) α 1 y ( s ) d s + u ( 0 ) + u ( 0 ) t = 0 t ( ( t s ) α 1 Γ ( α ) + a Γ ( α β ) ( Γ ( 2 β ) b ) ( 1 s ) α 1 ( 1 a ) Γ ( α ) Γ ( α β ) ( Γ ( 2 β ) b ) + b Γ ( α ) Γ ( 2 β ) ( a + t a t ) ( 1 s ) α β 1 ( 1 a ) Γ ( α ) Γ ( α β ) ( Γ ( 2 β ) b ) ) y ( s ) d s + t 1 ( a Γ ( α β ) ( Γ ( 2 β ) b ) ( 1 s ) α 1 ( 1 a ) Γ ( α ) Γ ( α β ) ( Γ ( 2 β ) b ) + b Γ ( α ) Γ ( 2 β ) ( a + t a t ) ( 1 s ) α β 1 ( 1 a ) Γ ( α ) Γ ( α β ) ( Γ ( 2 β ) b ) ) y ( s ) d s = 0 1 G ( t , s ) y ( s ) d s .

This completes the proof. □

Put k 1 = Γ ( α β ) ( Γ ( 2 β ) b ) + b Γ ( α ) Γ ( 2 β ) ( 1 a ) Γ ( α ) Γ ( α β ) ( Γ ( 2 β ) b ) and k 2 = a b Γ ( 2 β ) ( 1 a ) Γ ( α β ) ( Γ ( 2 β ) b ) . It is easy to check that the Green function G in the last result belongs to C ( [ 0 , 1 ] × [ 0 , 1 ] ) , G ( t , s ) > 0 for all ( t , s ) [ 0 , 1 ) × [ 0 , 1 ) ,
G ( t , s ) k 1 ( 1 s ) α β 1 1 and G ( t , s ) k 2 ( 1 s ) α β 1
for all ( t , s ) [ 0 , 1 ] × [ 0 , 1 ] . Consider the Banach space X = C 1 [ 0 , 1 ] with the norm x = max { x , x } and the cone
P = { x X : x ( t ) 0  and  x ( t ) 0  for all  t [ 0 , 1 ] } .
For each natural number n, define the operator Q n on P by
( Q n x ) ( t ) = 0 1 G ( t , s ) f n ( s , u ( s ) , u ( s ) , c D β u ( s ) ) d s .

Now, we prove that Q n is a completely continuous operator (see [2]).

Lemma 2.2 The operator Q n is a completely continuous operator.

Proof Let x P . Then, D β c x C [ 0 , 1 ] and D β c x 0 . Now, define ρ ( t ) = f n ( t , u ( t ) , u ( t ) , c D β u ( t ) ) for almost all t [ 0 , 1 ] . Then ρ L q [ 0 , 1 ] and ρ ( t ) m for almost all t [ 0 , 1 ] . By using the properties of fractional integral I α , it is easy to see that Q n x C [ 0 , 1 ] , ( Q n x ) ( t ) 0 and
( Q n x ) ( t ) = 1 Γ ( α 1 ) 0 t ( t s ) α 2 ρ ( s ) d s
for all t [ 0 , 1 ] . This implies that ( Q n x ) C [ 0 , 1 ] and ( Q n x ) 0 on [ 0 , 1 ] . Consequently, Q n maps P into P. In order to prove that Q n is a continuous operator, let x m be a convergent sequence in P and lim m x m = x . Thus, lim m x m ( j ) ( t ) = x ( j ) ( t ) uniformly on [ 0 , 1 ] for j = 0 , 1 . Since
D β c x ( t ) = 1 Γ ( 1 β ) d d t 0 t ( t s ) β ( x ( s ) x ( 0 ) ) d s = 1 Γ ( 1 β ) 0 t ( t s ) β x ( s ) d s ,
we get | c D β x m ( t ) c D β x ( t ) | x m x Γ ( 1 β ) 0 t ( t s ) β d s x m x Γ ( β ) and lim m c D β x m ( t ) = c D β x ( t ) uniformly on [ 0 , 1 ] . Also, we have | c D β x m ( t ) | x m Γ ( β ) on [ 0 , 1 ] , and so c D β x m x m Γ ( β ) . Now, put
ρ m ( t ) = f n ( t , x m ( t ) , x m ( t ) , c D β x m ( t ) ) and ρ ( t ) = f n ( t , x ( t ) , x ( t ) , c D β x ( t ) ) .
Then, it is easy to see that lim m ρ m ( t ) = ρ ( t ) for almost all t [ 0 , 1 ] , and there exists β L q [ 0 , 1 ] such that 0 ρ m ( t ) β ( t ) for almost all t [ 0 , 1 ] and all m 1 . Since f n is an L q -Caratheodory function, { x m } is bounded in C 1 [ 0 , 1 ] , and { c D β x m } is bounded in C [ 0 , 1 ] . Therefore, lim m ( Q n x m ) ( t ) = ( Q n x ) ( t ) uniformly on [ 0 , 1 ] . Since { ρ m } is L q -convergent on [ 0 , 1 ] ,
lim m ( Q n x m ) ( t ) = 1 Γ ( α 1 ) lim m 0 t ( t s ) α 2 ρ m ( s ) d s = ( Q n x ) ( t )
uniformly on [ 0 , 1 ] . Hence, Q n is a continuous operator. Now, we have to show that for each bounded sequence { x m } in P, the sequence { Q n x m } is relatively compact in C [ 0 , 1 ] . Choose a positive constant k such that x m k and x m k for all m. Note that c D β x m k Γ ( β ) and | 0 t ( t s ) α 2 ρ m ( s ) d s | ( 0 t ( t s ) ( α 2 ) p d s ) 1 p ( 0 t | ρ m ( s ) | q d s ) 1 q ( t d d ) 1 p ρ m q for all m, where d = ( α 2 ) p + 1 . But we have
0 ( Q n x m ) ( t ) = 0 1 G ( t , s ) ρ m ( s ) d s 0 1 G ( t , s ) β ( s ) d s β 1 Γ ( α )
and
0 ( Q n x m ) ( t ) = 1 Γ ( α 1 ) 0 t ( t s ) α 2 ρ m ( s ) d s 1 Γ ( α 1 ) 0 t ( t s ) α 2 β ( s ) d s 1 Γ ( α 1 ) ( 1 ( α 2 ) p + 1 ) 1 p β q
for all t [ 0 , 1 ] and m. This implies that { Q n x m } is bounded in C 1 [ 0 , 1 ] . Also, we have
| ( Q n x m ) ( t 2 ) ( Q n x m ) ( t 1 ) | = 1 Γ ( α 1 ) | 0 t 2 ( t 2 s ) α 2 ρ m ( s ) d s 0 t 1 ( t 1 s ) α 2 ρ m ( s ) d s | ρ m q Γ ( α 1 ) ( ( t 1 d + ( t 2 t 1 ) d t 2 d d ) 1 p + ( ( t 2 t 1 ) d d ) 1 p ) β q Γ ( α 1 ) ( ( t 1 d + ( t 2 t 1 ) d t 2 d d ) 1 p + ( ( t 2 t 1 ) d d ) 1 p )

for all 0 t 1 t 2 1 , where d = ( α 2 ) p + 1 . Hence, { ( Q n x m ) } is equicontinuous on [ 0 , 1 ] . Thus, { Q n x m } is relatively compact in C 1 [ 0 , 1 ] by the Arzela-Ascoli theorem. Hence, Q n is a completely continuous operator. □

We need the following result (see [2] and [21]).

Lemma 2.3 [21]

Let Y be a Banach space, P a cone in Y and Ω 1 and Ω 2 bounded open balls in Y centered at the origin with Ω ¯ 1 Ω 2 . Suppose that T : P ( Ω ¯ 2 Ω 1 ) P is a completely continuous operator such that T x x for all x P Ω 1 and T x x for all x P Ω 2 . Then T has a fixed point in P ( Ω ¯ 2 Ω 1 ) .

Theorem 2.4 For each natural number n, problem () has a solution u n P such that u n m k 2 α β , u n ( t ) m t α 1 Γ ( α ) and D β c u n ( t ) m t α β Γ ( α β + 1 ) for all t [ 0 , 1 ] .

Proof Let n 1 . It is sufficient to show that Q n has a fixed point u n in P with the desired conditions. In this way, note that
( Q n x ) ( t ) = 0 1 G ( t , s ) f n ( s , x ( s ) , x ( s ) , c D β x ( s ) ) d s m 0 1 G ( t , s ) d s m 0 1 k 2 ( 1 s ) α β 1 d s = m k 2 α β ,
and so Q n x Q n ( x ) m k 2 α β . Put Ω 1 = { x X : x < m k 2 α β } . Then Q n x x for all x P Ω 1 . If v n = h ( 1 n ) + r ( 1 n ) + k ( 1 n ) , then
| ( Q n x ) ( t ) | | 0 1 G ( t , s ) f n ( s , u ( s ) , u ( s ) , c D β u ( s ) ) d s | 0 1 | G ( t , s ) | ( v n + γ ( s ) w ( 1 + | x ( s ) | , 1 + | x ( s ) | , 1 + | c D β x ( s ) | ) ) d s k 1 ( v n + w ( 1 + x , 1 + x , 1 + c D β x ) γ 1 )
and
| ( Q n x ) ( t ) | = | 1 Γ ( α 1 ) 0 t ( t s ) α 2 f n ( s , x ( s ) , x ( s ) , c D β x ( s ) ) d s | 1 Γ ( α 1 ) 0 t ( t s ) α 2 ( v n + γ ( s ) w ( 1 + | x ( s ) | , 1 + | x ( s ) | , 1 + | c D β x ( s ) | ) ) d s 1 Γ ( α 1 ) ( v n t α 1 α 1 + w ( 1 + x , 1 + x , 1 + c D β x ) 0 t ( t s ) α 2 γ ( s ) d s )
for all x P and t [ 0 , 1 ] , because w is non-decreasing in all its variables. Since x x , x x , c D β x x Γ ( β ) x Γ ( β ) and
0 t ( t s ) α 2 γ ( s ) d s ( 1 d ) 1 / p γ q ,
where d = ( α 2 ) p + 1 , we have
Q n ( x ) k 1 ( v n + w ( 1 + x , 1 + x , 1 + x Γ ( β ) ) γ 1 )
and
( Q n x ) 1 Γ ( α 1 ) ( v n α 1 + w ( 1 + x , 1 + x , 1 + x Γ ( β ) ) ( 1 / d ) 1 / p γ q ) .
Hence, Q n x M ( v n α 1 + N w ( 1 + x , 1 + x , 1 + x Γ ( β ) ) ) , where N = max { γ 1 , ( 1 / d ) 1 / p γ q } and M = max { 1 Γ ( α 1 ) , k 1 } . Since
lim v 1 v w ( 1 + v , 1 + v , 1 + v ) = 0 ,
there exists a positive constant L such that
M ( v n α 1 + N w ( 1 + v , 1 + v , v Γ ( β ) ) ) < v
for all v L . Thus, Q n x < x for all x P with x L . Put Ω 2 = { x X : x < L } . Then Q n x < x for all x P Ω 2 . By using last result, Q n has a fixed point u n in P ( Ω ¯ 2 Ω 1 ) . But u n = ( Q n u n ) ( t ) m k 2 α β and
( Q n x ) ( t ) = 1 Γ ( α 1 ) 0 t ( t s ) α 2 f n ( s , x ( s ) , x ( s ) , c D β x ( s ) ) d s m Γ ( α 1 ) 0 t ( t s ) α 2 d s = m t α 1 Γ ( α )
for all t [ 0 , 1 ] and x P . Since 0 t ( t s ) β s α 1 d s = Γ ( α ) Γ ( 1 β ) Γ ( α β + 1 ) t α β ,
D β c u n ( t ) = 1 Γ ( 1 β ) 0 t ( t s ) β u n ( s ) d s m Γ ( α ) Γ ( 1 β ) 0 t ( t s ) β s α 1 d s = m t α β Γ ( α β + 1 )

for all t [ 0 , 1 ] . This completes the proof. □

Now, we give our last result.

Theorem 2.5 Problem () has a solution u such that u ( t ) m k 2 α β , u ( t ) m t α 1 Γ ( α ) and D β c u ( t ) m t α β Γ ( α β + 1 ) for all t [ 0 , 1 ] .

Proof By using Theorem 2.4, one gets that for each natural number n, problem () has a solution u n P with the desired conditions. Thus, h ( u n ( t ) ) h ( m k 2 α β ) , r ( | u n ( t ) | ) r ( m t α 1 Γ ( α ) ) and k ( | c D β u n ( t ) | ) k ( m t α β Γ ( α β + 1 ) ) for all t [ 0 , 1 ] and n. Also, we have c D β u n u n Γ ( β ) . Suppose that
S ( t ) = h ( m k 2 α β ) + r ( m t α 1 Γ ( α ) ) + k ( m t α β Γ ( α β + 1 ) ) .
Then
m f n ( t , u n ( t ) , u n ( t ) , c D β u n ( t ) ) S ( t ) + γ ( t ) w ( 1 + u n , 1 + u n , 1 + c D β u n ) S ( t ) + γ ( t ) w ( 1 + u n , 1 + u n , 1 + u n Γ ( β ) )
for almost all t [ 0 , 1 ] and n. Since 0 G ( t , s ) k 1 , we get
0 u n ( t ) = 0 1 G ( t , s ) f n ( s , u n ( s ) , u n ( s ) , c D β u n ( s ) ) d s k 1 ( 0 1 S ( s ) d s + w ( 1 + u n , 1 + u n , 1 + u n Γ ( β ) ) γ 1 )
and
0 u n ( t ) 1 Γ ( α 1 ) ( 0 t ( t s ) α 2 S ( s ) d s + w ( 1 + u n , 1 + u n , 1 + u n Γ ( β ) ) 0 t ( t s ) α 2 γ ( s ) d s ) .
We show that 0 t ( t s ) α 2 S ( s ) d s is bounded on [ 0 , 1 ] . Let d = ( α 2 ) p + 1 . Note that
0 1 ( t s ) α 2 h ( m k 2 α β ) d s = h ( m k 2 α β ) 0 1 ( t s ) α 2 d s = 1 α 1 h ( m k 2 α β ) = : η 1 < , 0 t ( t s ) α 2 r ( m s α 1 Γ ( α ) ) d s = ( 1 d ) 1 / p ( Γ ( α ) m ) 1 ( α 1 ) q ( 0 ( m Γ ( α ) ) 1 α 1 r q ( s α 1 ) d s ) 1 / q = : η 2 <
and
0 t ( t s ) α 2 k ( m s α β Γ ( α β + 1 ) ) d s = ( 1 d ) 1 / p ( Γ ( α β + 1 ) m ) 1 ( α β ) q ( 0 ( m Γ ( α β + 1 ) ) 1 α β k q ( s α β ) d s ) 1 / q = : η 3 < .
Thus, 0 t ( t s ) α 2 S ( s ) d s η for all t [ 0 , 1 ] , where η = η 1 + η 2 + η 3 . Also, we have
0 1 S ( s ) d s 1 α 1 h ( m k 2 α β ) + ( Γ ( α ) m ) 1 α 1 0 ( m Γ ( α ) ) 1 α 1 r ( s α 1 ) d s + ( Γ ( α β + 1 ) m ) 1 α β 0 ( m Γ ( α β + 1 ) ) 1 α β k ( s α β ) d s < .
Since
u n = k 1 ( 0 1 S ( s ) d s + w ( 1 + u n , 1 + u n , 1 + u n Γ ( β ) ) γ 1 )
and
u n 1 Γ ( α 1 ) ( η + w ( 1 + u n , 1 + u n , 1 + u n Γ ( β ) ) ( 1 d ) 1 / p γ q ) ,
we get u n M ( Φ + K w ( 1 + u n , 1 + u n , 1 + u n Γ ( β ) ) ) for all n, where Φ = max { η , 0 1 S ( s ) d s } , K = max { γ 1 , ( 1 d ) 1 / p γ q } and also M = max { k 1 , 1 Γ ( α 1 ) } . On the other hand, there exists a positive constant L such that M ( Φ + K w ( 1 + v , 1 + v , 1 + v Γ ( β ) ) ) < v for all v L , and so u n < L for all n. Thus, for almost all t [ 0 , 1 ] and all n, we have f n ( t , u n ( t ) , u n ( t ) , c D β u n ( t ) ) R ( t ) , where
R ( t ) = S ( t ) + γ ( t ) w ( 1 + L , 1 + L , 1 + L Γ ( β ) ) .
Note that R L q [ 0 , 1 ] . We show that { u n } is equicontinuous on [ 0 , 1 ] . Let ρ n ( t ) = f n ( t , u n ( t ) , u n ( t ) , c D β u n ( t ) ) and 0 t 1 < t 2 T . Then
| u n ( t 2 ) u n ( t 1 ) | = 1 Γ ( α 1 ) | 0 t 2 ( t 2 s ) α 2 ρ n ( s ) d s 0 t 1 ( t 1 s ) α 2 ρ n ( s ) d s | 1 Γ ( α 1 ) ( 0 t 1 ( ( t 1 s ) α 2 ( t 2 s ) α 2 ) ρ n ( s ) d s + t 1 t 2 ( t 2 s ) α 2 ρ n ( s ) d s ) 1 Γ ( α 1 ) ( 0 t 1 ( ( t 1 s ) α 2 ( t 2 s ) α 2 ) R ( s ) d s + t 1 t 2 ( t 2 s ) α 2 R ( s ) d s ) ,
and so
| u n ( t 2 ) u n ( t 1 ) | R q Γ ( α 1 ) ( ( t 1 d + ( t 2 t 1 ) d t 2 d d ) 1 / p + ( ( t 2 t 1 ) d d ) 1 / p ) .
Hence, { u n } is equicontinuous on [ 0 , 1 ] . Since { u n } is a bounded sequence in C [ 0 , 1 ] , by using the Arzela-Ascoli theorem, without loss of generality, we can assume that { u n } is convergent in C [ 0 , 1 ] . Let lim n u n = u . Then, it is easy to see that D β c u n ( t ) = 1 Γ ( α 1 ) 0 t ( t s ) β u n ( s ) d s , and D β c u n ( t ) uniformly converges to 1 Γ ( α 1 ) 0 t ( t s ) β u ( s ) d s on [ 0 , 1 ] . Thus, D β c u n converges to D β c u in C [ 0 , 1 ] . Hence,
lim n f n ( t , u n ( t ) , u n ( t ) , c D β u n ( t ) ) = f ( t , u ( t ) , u ( t ) , c D β u ( t ) )
for almost all t [ 0 , 1 ] . Since R L q [ 0 , 1 ] , by using the dominated convergence theorem on the relation
u n ( t ) = 0 1 G ( t , s ) f n ( s , u n ( s ) , u n ( s ) , c D β u n ( s ) ) d s ,

we get u ( t ) = 0 1 G ( t , s ) f ( s , u ( s ) , u ( s ) , c D β u ( s ) ) for all t [ 0 , 1 ] . This completes the proof. □

2.1 Examples for the problem

Example 2.1 Let ρ 1 , ρ 2 L q [ 0 , 1 ] , ρ 1 ( t ) m > 0 for almost all t in [ 0 , 1 ] . Suppose that
f ( t , x , y , z ) = ρ 1 ( t ) + 1 x 1 3 λ + 1 y 1 4 + 1 z 1 4 + | ρ 2 ( t ) | ( x 1 3 + y 1 4 + z 1 4 )

on [ 0 , 1 ] × B , λ = ( a u ( 1 ) ) 1 3 , h ( x ) = 1 x 1 3 λ whenever x 1 3 λ 0 and h ( x ) = 0 whenever x 1 3 λ < 0 , r ( x ) = 1 x 1 4 , k ( x ) = 1 x 1 4 , w ( x , y , z ) = ( x 1 3 + y 1 4 + z 1 4 + 1 ) and γ ( t ) = ρ 1 ( t ) + | ρ 2 ( t ) | . Then Theorem 2.5 guarantees that problem () has a positive solution.

Example 2.2 Consider the nonlinear mixed problem of singular fractional boundary value problem
D 7 3 u ( t ) = t = 1 = 1 u ( t ) 1 3 - ρ = 1 u ( t ) 1 4 = 1 ( c D 1 4 u ( t ) ) 1 4 = 2 ( u ( t ) 1 3 = u ( t ) 1 4 = ( c D 1 4 u ( t ) ) 1 4 = 1 ) )

via boundary value conditions u ( 0 ) = 1 4 u ( 1 ) , u ( 0 ) = 1 3 ( c D 1 4 ) u ( 1 ) and u ( 0 ) = u ( 0 ) = = u ( n 1 ) ( 0 ) = 0 , where ρ = ( ( 1 4 ) u ( 1 ) ) 1 3 . Let f

Then the map f is singular at t = 0 , and f satisfies the desired conditions, where h ( x ) = 1 x 1 3 ρ whenever x 1 3 ρ 0 and h ( x ) = 0 whenever x 1 3 ρ < 0 , r ( x ) = 1 x 1 4 , k ( x ) = 1 x 1 4 , w ( x , y , z ) = x 1 3 + y 1 4 + z 1 4 + 1 , ρ 1 ( t ) = t + 1 > 1 = m , ρ 2 ( t ) = 2 and γ ( t ) = ρ 1 ( t ) + | ρ 2 ( t ) | . Then Theorem 2.5 guarantees that this problem has a positive solution.

3 Conclusions

One of the most interesting branches is obtaining solutions of singular fractional differential via boundary value problems. Having these things in mind, we study the existence of solutions for a singular nonlinear fractional boundary value problem. Two illustrative examples illustrate the applicability of the proposed method. It seems that the obtained results could be extended to more general functional spaces. Finally, note that all calculations in proofs of the results depend on the definition of the fractional derivative.

Declarations

Acknowledgements

Research of the second and third authors was supported by Azarbaijan Shahid Madani University. Also, the authors express their gratitude to the referees for their helpful suggestions, which improved the final version of this paper.

Authors’ Affiliations

(1)
Department of Chemical and Materials Engineering, Faculty of Engineering, King Abdulaziz University
(2)
Department of Mathematics, Cankaya University
(3)
Institute of Space Sciences
(4)
Department of Mathematics, Azarbaijan Shahid Madani University

References

  1. Bai Z, Lu H: Positive solutions for boundary value problem of nonlinear fractional differential equation. J. Math. Anal. Appl. 2005, 311: 495-505. 10.1016/j.jmaa.2005.02.052MathSciNetView ArticleGoogle Scholar
  2. Baleanu D, Mohammadi H, Rezapour S: Positive solutions of a boundary value problem for nonlinear fractional differential equations. Abstr. Appl. Anal. 2012., 2012: Article ID 837437Google Scholar
  3. Baleanu D, Mohammadi H, Rezapour S: Some existence results on nonlinear fractional differential equations. Philos. Trans. R. Soc. Lond. A 2013., 371: Article ID 20120144Google Scholar
  4. Baleanu D, Mohammadi H, Rezapour S: On a nonlinear fractional differential equation on partially ordered metric spaces. Adv. Differ. Equ. 2013., 2013: Article ID 83Google Scholar
  5. Kilbas AA, Trujillo JJ: Differential equations of fractional order: methods, results and problems. Appl. Anal. 2002, 81(2):435-493. 10.1080/0003681021000022032MathSciNetView ArticleGoogle Scholar
  6. Zhang SQ: Existence of positive solution for some class of nonlinear fractional differential equations. J. Math. Anal. Appl. 2003, 278: 136-148. 10.1016/S0022-247X(02)00583-8MathSciNetView ArticleGoogle Scholar
  7. Agarwal RP, O’Regan D, Stanek S: Positive solutions for mixed problems of singular fractional differential equations. Math. Nachr. 2012, 285(1):27-41. 10.1002/mana.201000043MathSciNetView ArticleGoogle Scholar
  8. Cabrera IJ, Harjani J, Sadarangani KB: Existence and uniqueness of positive solutions for a singular fractional three-point boundary value problem. Abstr. Appl. Anal. 2012., 2012: Article ID 803417Google Scholar
  9. Cao J, Chen H: Positive solution of singular fractional differential equation in Banach space. J. Appl. Math. 2011., 2011: Article ID 352341Google Scholar
  10. Liu Y, Nieto JJ, Otero-Zarraquinos O: Existence results for a coupled system of nonlinear singular fractional differential equations with impulse effects. Math. Probl. Eng. 2013., 2013: Article ID 498781Google Scholar
  11. Wei X, Lu X: The periodic solutions of the compound singular fractional differential system with delay. Int. J. Differ. Equ. 2010., 2010: Article ID 509286Google Scholar
  12. Zhou WX, Chu YD, Baleanu D: Uniqueness and existence of positive solutions for a multi-point boundary value problem of singular fractional differential equations. Adv. Differ. Equ. 2013., 2013: Article ID 114Google Scholar
  13. Baleanu D, Agarwal RP, Mohammadi H, Rezapour S: Some existence results for a nonlinear fractional differential equation on partially ordered Banach spaces. Bound. Value Probl. 2013., 2013: Article ID 112Google Scholar
  14. Baleanu D, Bhrawy AH, Taha TM: Two efficient generalized Laguerre spectral algorithms for fractional initial value problems. Abstr. Appl. Anal. 2013., 2013: Article ID 546502Google Scholar
  15. Baleanu D, Diethelm K, Scalas E, Trujillo JJ Series on Complexity, Nonlinearity and Chaos. In Fractional Calculus Models and Numerical Methods. World Scientific, Singapore; 2012.Google Scholar
  16. Bhrawy AH, Alghamdi MA: A shifted Jacobi-Gauss-Lobatto collocation method for solving nonlinear fractional Langevin equation involving two fractional orders in different intervals. Bound. Value Probl. 2012., 2012: Article ID 62Google Scholar
  17. Bhrawy AH, Alofi AS: The operational matrix of fractional integration for shifted Chebyshev polynomials. Appl. Math. Lett. 2013, 26: 25-31. 10.1016/j.aml.2012.01.027MathSciNetView ArticleGoogle Scholar
  18. Bhrawy AH, Al-Shomrani MM: A shifted Legendre spectral method for fractional-order multi-point boundary value problems. Adv. Differ. Equ. 2012., 2012: Article ID 8Google Scholar
  19. Bhrawy AH, Tharwata MM, Yildirim A: A new formula for fractional integrals of Chebyshev polynomials: application for solving multi-term fractional differential equations. Appl. Math. Model. 2013, 37: 4245-4252. 10.1016/j.apm.2012.08.022MathSciNetView ArticleGoogle Scholar
  20. Podlubny I Mathematics in Science and Engineering 198. In Fractional Differential Equations. Academic Press, San Diego; 1999.Google Scholar
  21. Krasnoselski MA: Positive Solutions of Operator Equations. Noordhoff, Groningen; 1964.Google Scholar

Copyright

© Baleanu et al.; licensee Springer. 2013

This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.