Mathematical Computation June 2013, Volume 2, Issue 2, PP.13-18
The Existence of Solutions for a Third-Order Multi-Point Boundary Value Problem at Resonance* Weihua Jiang#, Jiqing Qiu, Ruiyan Li College of Sciences, Hebei University of Science and Technology, Shijiazhuang, 050018, Hebei, P. R. China #Email: weihuajiang@hebust.edu.cn
Abstract By means of the coincidence degree theory due to Mawhin, suitable Banach space has been constructed and appropriate operators have been defined, the solutions to a third-order multi-point boundary value problem at resonance have been obtained, in which dim KerL 3 . Keywords: Resonance; Fredholm Operator; Multi-Point Boundary Value Problem; Coincidence Degree Theory
1
INTRODUCTION
In this paper, the solutions to the third-order multi-point boundary value problem have been studied:
u(t ) f t , u(t ), u(t ), u(t ) e(t ), t (0,1), m
n
l
i 1
j 1
k 1
(1.1)
u (0) i u (i ) , u (0) i u ( j ) , u (1) k u ( k ) , where f :[0,1] R3 R is a 0 1 2 n 1 , 0 1 2
Caratheodory l 1 .
function,
e(t ) L1[0,1],
0 1 2
(1.2) m 1
,
The boundary value problem (1.1)-(1.2) is a problem at resonance if Lu : u(t ) 0 has non-trivial solutions under the boundary condition (1.2), i.e. dim KerL 1 . The solutions to first-order, second-order and high-order multi-point boundary value problems at resonance have been studied previously (see, for example [1-5]), in which dim KerL 1 . In [6-8], the second-order multi-point boundary value problems at resonance have been discussed when dim KerL 2 . Motivated by the above results, we will investigate the solutions to the problem (1.1)-(1.2) with dim KerL 3 . To the best of our knowledge, this is the first paper to study the resonance problems with dim KerL 3 . In this section, the necessary background definitions and the key theorem duo to Mawhin are provided. In section 2, the main results of the problem (1.1)-(1.2) will be stated and proved. Let Y and Z be real Banach spaces and let L : domL Y Z be a Fredholm operator with index zero, P : Y Y , Q : Z Z be projectors such that
Im P KerL , KerQ Im L , Y KerL KerP , Z Im L Im Q . It follows that
L
domL KerP
: domL
KerP Im L
is invertible. The inverse is defined by K P . *
MSC: 34B10, 34B15 This work is supported by the Natural Science Foundation of China (11171088) and the Natural Science Foundation of Hebei Province (A2013208108) - 13 www.ivypub.org/mc
If is an open bounded subset of Y , domL
, the map N : Y Z will be called L-compact on if
QN is bounded and K P ( I Q) N : Y is compact. Theorem 2.1 [9]. Let L : domL Y Z be a Fredholm operator of index zero and N : Y Z L-compact on . It is assumed that the following conditions are satisfied: (1) Lu Nu for every (u, ) domL \ KerL (2) Nu Im L for every u KerL (3) deg QN
KerL
(0,1) ;
;
KerL, 0 0 , where Q : Z Z is a projection such that Im L KerQ .
,
Then the equation Lu Nu has at least one solution in domL
2
.
MAIN RESULTS
Let Y C 2 [0,1] be a real Banach space with norm u max u , u , u , where u 1 Z L1[0,1] with norm u 1 u (t ) dt . L : domL Y Z is defined by 0
max u(t ) and let t[0,1]
Lu u, u domL,
n l where domL u Y : u (0) i u (i ), u (0) j u ( j ), u (1) k u ( k ), u Z . i 1 j 1 k 1 m
A nonlinear operator N : Y Z is defined by Nu(t ) f t , u(t ), u(t ), u(t ) e(t ), t [0,1].
Then boundary value problem (1.1)-(1.2) can be written by Lu Nu , u domL . In this paper, it is supposed that the following conditions hold: m
( C1 )
n
l
i 1
i
j
j 1
k 1
k
1,
m
m
i i
i 1
Q11 Q21 Q31
( C2 ) △ Q1t Q2t Q31 Q1t
2
Q2t
2
Q3t
i i
i 1
n
2
j j 0 ; j 1
a11
a12
a13
: a21
a22
a23 0 .
a31
a32
a33
2
where Qi : Z R, i 1, 2,3 are defined as follows: m
Q1 y i i 1
i 0
n
j
l
(i s)2 y(s)ds , Q2 y j ( j s) y (s )ds , Q3 y k y (s)ds j 1
0
k 1
1
k
Lemma 2.1 Supposing that conditions ( C1 ) and ( C2 ) hold, then L : domL Y Z is a Fredholm operator of index zero. Furthermore, the linear continuous operator Q : Z Z can be defined by
Qy T1 y T2 y t T3 y t 2 . where
T1 y
1 △11Q1 y △12Q2 y △13Q3 y , △
T2 y
1 △21Q1 y △22Q2 y △23Q3 y , △
1 △31Q1 y △32Q2 y △33Q3 y . △ △ij is the algebraic cofactor of aij , i 1, 2,3, j 1, 2,3. T3 y
Projection P : Y Y is defined by - 14 www.ivypub.org/mc
1 Pu (t ) u (0) u (0)t u (0)t 2 , t [0,1]. 2 The linear operator K P : Im L domL be written by
KerP , which is the inverse of L
KP y
1 t (t s)2 y(s)ds, 2 0
domL KerP
: domL
KerP Im L , can
y Im L.
It is clear that
KP y y 1 ,
y Z.
(2.1)
Proof Obviously, KerL a bt ct 2 | a, b, c R . Now, it is shown that Im L y | y Z , Q1 y Q2 y Q3 y 0.
In fact, if Lu(t ) y(t ), u domL , then
1 1 t u (t ) u (0) u (0)t u (0)t 2 (t s)2 y(s)ds . 2 2 0 By (1.2), Q1 y Q2 y Q3 y 0 can be obtained. On the other hand, if Q1 y Q2 y Q3 y 0 , take
u (t )
1 t (t s)2 y(s)ds. 2 0
Then u(t ) y(t ) and u satisfies condition (1.2). For y Z , considering the definition of △ and △ij , we have 1 T1 T1 y △11Q1 T1 y △12Q2 T1 y △13Q3 T1 y △ 1 △11Q11 △12Q21 △13Q31 T1 y △ T1 y .
1 △11Q1 T2 y t △12Q2 T2 y t △13Q3 T2 y t △ 1 △11Q1 t △12Q2 t △13Q3 t T2 y △ 0.
T1 T2 y t
T1 T3 y t 2
1 △11Q1 T3 y t 2 △12Q2 T3 y t 2 △13Q3 T3 y t 2 △ 1 △11Q1 t 2 △12Q2 t 2 △13Q3 t 2 T3 y △ 0.
Similarly,
it
can
be
noted
that
T2 T1 y T2 T3 y t 2 T3 T1 y T3 T2 y t 0 , T2 T2 y t T2 y ,
T3 T3 y t 2 T3 y . Thus, Q2 y Qy is obtained, meaning that Q : Z Z is a projector operator. Now, it will be proved that KerQ Im L . Obviously, Im L KerQ . If y KerQ , then Qy 0 , thus T1 y T2 y T3 y 0 , i.e. △11Q1 y △12Q2 y △13Q3 y 0 , △21Q1 y △22Q2 y △23Q3 y 0 , △31Q1 y △32Q2 y △33Q3 y 0 . - 15 www.ivypub.org/mc
△11 △12 △13 Since △21 △22 △23
△31 △32 △33 For y Z ,
1 0 , we get Q1 y Q2 y Q3 y 0 , i.e. y Im L . So, KerQ Im L . △2
y Qy ( y Qy) , by
y Qy KerQ Im L
and Qy Im Q , we get Z Im L
ImQ .
Take y Im L Im Q . By y Im L , we have Q1 y Q2 y Q3 y 0 . Considering y Im Q , we get y Qy 0 . So, Z Im L Im Q . Since dim KerL co dimIm L dimIm Q 3 , it can be said that L is a Fredholm operator of index zero. Now, it is shown that the generalized inverse K P : Im L domL
KerP can be written by
1 t (t s)2 y (s)ds . 2 0 In fact, for y Im L , we have LK P y K P y y . For u domL KerP , we get KP y
1 t 1 (t s)2 u (s)ds u (0) u (0)t u (0)t 2 u (t ). 0 2 2 Considering u KerP , we get u(0) u(0) u(0) 0. Thus, K P Lu u. So, we get K P Lu
KP L
domL KerP
1
.
The following theorem is our main results. Theorem 2.1 Suppose ( C1 ), ( C2 ) and the following conditions hold: ( H1 ) functions , , , , L1[0,1] and a constant [0,1) exist such that, for all x1 , x2 , x3 R, t [0,1] , one of the following conditions holds:
f t , x1 , x2 , x3 (t ) (t ) x1 (t ) x2 (t ) x3 (t ) xi , i 1, 2,3.
Where
1 4
1 1 1 .
( H 2 ) a constant A 0 exists such that if u (t ) A or u(t ) A or u(t ) A , for all t [0,1] , one of the following inequality holds:
Qi Nu (t )
0 , t [0,1], i 1, 2,3.
( H 3 ) a constant B 0 exists such that, for any a, b, c R satisfying a2 b2 c2 B , either (1) aT1 N a bt ct 2 bT2 N a bt ct 2 cT3 N a bt ct 2 0, t [0,1] or (2) aT1 N a bt ct 2 bT2 N a bt ct 2 cT3 N a bt ct 2 0, t [0,1] Then the problem (1.1)-(1.2) has at least one solution. Proof This theorem will be verified by the following three steps. Step 1 Set 1 u domL \ KerL : Lu Nu, for some [0,1]
It is proved that 1 is bounded. In fact, u 1 means Nu Im L . Therefore, we get Q1 Nu Q2 Nu Q3 Nu 0. By ( H 2 ), ti [0,1], i 1, 2,3 exists such that u(t0 ) A , u(t1 ) A , u(t2 ) A . It follows from t0
u (0) u (t0 ) u (s)ds A u 0
- 16 www.ivypub.org/mc
,
t1
u(t ) u(t1 ) u (s)ds A u t
, t [0,1]
and t2
u(t ) u (t2 ) u (s)ds A u 1 , t [0,1] t
that Pu max Pu ,
Pu
,
Pu
6 A 3 Nu 1 .
In addition, ( I P)u K P L( I P)u L( I P)u 1 Lu 1 Nu 1 .
Therefore, u Pu ( I P)u Pu (I P)u 6 A 4 Nu 1 .
This, together with ( H1 ), implies
u 6A 4 1
1
u
1
u
1
u
1
u
e1
So, we have u
1 1 4 1 1
1
6A 4
1
4 e14
1
u
.
Since [0,1) , we get that there exists a constant M 0 such that u M , i.e. 1 is bounded. Step 2 Let 2 u KerL : Nu Im L.
u 2 means that constants a, b, c R exist such that u(t ) a bt ct 2 and Qi N a bt ct 2 0, i 1, 2,3. It follows from ( H 3 ) that u 3 B , i.e. 2 is bounded.
Step 3 A linear isomorphism J : KerL Im Q is defined by
J a bt ct 2
1 △11a △12b △13c △21a △22b △23c t △31a △32b △33c t 2 △
It is assumed that ( H 3 )(1) holds. Take
3 u KerL : Ju (1 )QNu 0, for some [0,1] . It will be verified that 3 is bounded. For u 3 , we have u(t ) a bt ct 2 and
J a bt ct 2 QN a bt ct 2 . Thus
△11a △12b △13c (1 )T1 N a bt ct 2 ,
△21a △22b △23c (1 )T2 N a bt ct 2 , △31a △32b △33c (1 )T3 N a bt ct 2 .
If 1 , then a b c 0 . If 0 , then Qi N a bt ct 2 0 , by ( H 3 ), u 3 B . If (0,1) , it is assumed that a2 b2 c2 B , by ( H 3 ) (1), we get
a 2 b2 c2 (1 ) aT1 N a bt ct 2 bT2 N a bt ct 2 cT3 N a bt ct 2 0 a contradiction. So, a2 b2 c2 B , i.e. 3 is bounded. - 17 www.ivypub.org/mc
Remark 2.1 If ( H 3 ) (2) holds, then take
3 u KerL : Ju (1 )QNu 0, for some [0,1] . Similarly, we can get that 3 is bounded. Then, it is shown that all the conditions of Theorem 1.1 are satisfied. Take an open bounded subset Y
3
i
{0} . By ( H1 ), it can be inferred that QN
i 1
is bounded. By
Arzela- Ascoli theorem, K p ( I Q) N : Y that is compact is acquired. Thus, N is L-compact on . It follows from Step 1 and Step 2 that (1) Lu Nu for every (u, ) domL \ KerL (2) Nu Im L , for every u KerL It will be further shown that (3) deg QN
KerL
,
(0,1);
. .
KerL, 0 0.
Let
H (u, ) Ju (1 )QNu, [0,1].
It follows from Step 3 that H (u, ) 0 , u KerL , [0,1] . Therefore, by the homotopy degree, we get deg QN KerL , KerL, 0 deg H , 0 , KerL, 0
deg H ,1 ,
KerL, 0 deg J ,
property of
KerL, 0 1 0.
By Theorem 1.1, it can be obtained that Lu Nu has at least one solution in domL one solution in Y . The proof is completed.
, i.e. (1.1)-(1.2) has at least
REFERENCES [1]
W. Feng, J. R. L. Webb. Solvability of $m$-point boundary value problems with nonlinear growth, J. Math. Anal. Appl. 212(1997), 467-480.
[2]
B. Liu. Solvability of multi-point boundary value problem at resonance-Part Ⅳ, Appl. Math. Comput. 143(2003) 275-299.
[3]
R. Ma. Existence results of a $m$-point boundary value problem at resonance,J. Math. Anal. Appl. 294(2004), 147-157.
[4]
C.P.Gupta. On a third-order boundary value problem at resonance, Diff, Integral Equat, 2(1989), 1-12.
[5]
S.Lu, W.Ge. On the existence of $m$-point boundary value problem at resonance for higher order differential equation, J. Math. Anal. Appl. 287(2003), 522-539.
[6]
N. Kosmatov. Multi-point boundary value problems on an unbounded domain at resonance, NonlinearAnal., 68(2008), 2158-2171.
[7]
X. Zhang, M. Feng. W. Ge. Existence result of second-order differential equations with integral boundary conditions at resonance, J. Math. Anal. Appl.353 (2009), 311-319.
[8]
Z. Du, F. Meng. Solutions to a second-order multi-point boundary value problem at resonance, Acta Math. Scientia, 30B (5) (2010), 1567-1576.
[9]
J.Mawhin. Topological degree methods in nonlinear boundary value problems, in:NSFCBMS Regional Conference Series in Mathematics, American Mathematical Society, Providence, RI, 1979.
AUTHORS 1Weihua
Jiang was born in Hebei, China, in 1964. She received
the B.S. degree in mathematics from Xi'an Jiaotong University, Xi'an, China, in 1991 and the Ph.D. degree in mathematics from Hebei Normal University, Shijiazhuang, China, in 2009. Her research interests are in boundary value problems and functional
2Jiqing
Qiu (1956-): Male, Han nationality, Ph.D., professor.
His research interests are in Rrobust control. 3Ruiyan
Li (1983-): Female, Han nationality, master. Research
direction: Boundary value problems.
analysis. - 18 www.ivypub.org/mc