Journal Pre-proof

Existence of solution of a three-point boundary value problem via

variational approach

Wen Lian, Zhanbing Bai, Zengji Du

PII: S0893-9659(20)30076-8
DOI: https://doi.org/10.1016/j.aml.2020.106283
Reference: AML 106283

To appear in: Applied Mathematics Letters

Received date : 11 December 2019

Revised date : 6 February 2020
Accepted date : 6 February 2020

Please cite this article as: W. Lian, Z. Bai and Z. Du, Existence of solution of a three-point
boundary value problem via variational approach, Applied Mathematics Letters (2020), doi:
https://doi.org/10.1016/j.aml.2020.106283.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the
addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive
version of record. This version will undergo additional copyediting, typesetting and review before it
is published in its final form, but we are providing this version to give early visibility of the article.
Please note that, during the production process, errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier Ltd.

 Journal Pre-proof
*Manuscript
Click here to view linked References

of
Existence of solution of a three-point boundary value
problem via variational approach✩

pro
Wen Liana , Zhanbing Baia,∗, Zengji Dub
a College of Mathematics and Systems Science, Shandong University of Science and
Technology, Qingdao 266590, PR China
b School of Mathematics and Statistics, Jiangsu Normal University, Xuzhou 221116, PR

China

Abstract
re-
By the use of a constructing adequate real functional and choosing an ap-
propriate admissible function space, the existence and multiplicity of solutions
to a three-point second-order differential system is proved via Mountain Pass
Lemma. The interesting point is the boundary value conditions are imposed on
admissible space rather than the functionals.
Keywords: Three-point BVPs; Critical point; Variational approach;
P
Differential system.
2010 MSC: 34B15 34A08

1. Introduction
al

The variational approach, together with the critical point theory, is one of the
important methods in the study of two-point boundary value problems (BVPs)
of ordinary differential equation [3, 15]. The approach is also much effective in
the research of boundary value problems of differential systems [6, 13, 16].
A basic problem in mechanics (classical and celestial) is the study of the
rn

periodic solutions of Hamiltonian systems

J u̇(t) + ∇H(t, u(t)) = 0, (1.1)
where ∇H(t, u) = Du H(t, u). Mawhin and Willem [13] pointed out that the
variational structure of the problem suggests that the best results should be
Jou

obtained through a variational approach.

To second-order Hamiltonian systems, Mawhin and Willem [13] studied the
periodic solutions of convex Hamiltonian system

q̈ + ∇H(t, q(t)) = 0, t ∈ (0, T ),
(1.2)
q(0) − q(T ) = q̇(0) − q̇(T ) = 0

✩ This work is supported by NSFC(11571207, 11871251) and the Taishan Scholar project.
∗ Corresponding author
Email addresses: sdkjdx18lian@163.com (Wen Lian), zhanbingbai@163.com (Zhanbing
Bai), duzengji@163.com (Zengji Du)

Preprint submitted to AML February 6, 2020

 Journal Pre-proof

of
where H : [0, T ] × RN → R is measurable in t for every q ∈ Rn and con-
tinuously differentiable and convex in q for a.e. t ∈ [0, T ]. If there exist
l ∈ L4 ([0, T ]; Rn ), γ ∈ L2 ([0, T ]; Rn ) and α ∈ (0, 2π/T ) such that
α 2
(l(t), u) ≤ H(t, u) ≤ |u| + γ(t),

pro
2
and the coercivity on the kernel
Z T
H(t, u)dt → ∞, as |u| → ∞,
0

then Problem (1.2) has at least one solution.

Tian et. al. [16] discussed the following problem
 d ′
dt ϕp (u (t)) + ∇F (t, u(t)) = 0,
re-
u(0) − u(T ) = u′ (0) − u′ (T ) = 0,

where ϕp (x) = |x|p−2 x for x ∈ Rn . An existence theorem of periodic solutions

(1.3)

was obtained under the coercivity condition and

p−2 α2 p
(l(t), |u| 2 u) ≤ F (t, u) ≤ |u| + γ(t).
p
P
The result extended that given by Mawhin and Willem [13].
In applications to differential equations, critical points correspond to weak
solutions of the equation. In deed this fact makes critical point theory an
important existence tool in studying differential equations with Dirichlet and
periodic boundary conditions [13]. To Sturm-Liouville boundary conditions,
Tian et. al. have studied many BVPs with space H 1 ([0, 1]). The key point is
al

constructing adequate functionals. For examples [17], to BVPs


(p(t)u′ )′ − q(t)u + f (t, u) = 0, 0 < t < 1;
αu(0) − βu′ (0) = 0, γu(1) + σu′ (1) = 0,
rn

setting
Z 1 Z 1
1
Φ(u) = [p(t)|u′ (t)|2 + q(t)|u(t)|2 ]dt − F (t, u(t))dt
2 0 0
σp(1) βp(0)
+ |u(1)|2 + |u(0)|2 .
2γ 2α
Jou

Recently, many methods, such as fixed-point theory, topology method, upper

and lower solution method, have been used to study the solutions of nonlocal
boundary value problems, see [2, 4, 5, 9–12, 18] and numerous works follows.
However, to my best knowledge, there is no works study multi-point boundary
value problems using variational method.
In this paper, we are going to study the existence of multiple solutions to
the following three-point differential system boundary value problem

(P (t)x′ (t))′ + f (t, x(t)) = 0, a.e. 0 < t < 1,
(1.4)
x(0) = 0, x(1) = βx(η).
Rx
Setting F (t, x) = 0 f (t, u)du, we will use following assumptions:

2
 Journal Pre-proof

of
(H1) P : [0, 1] → Rn×n is a continuously symmetric matrix, {pi (t)} is the
eigenvalue of P (t) and 0 < a ≤ min min pj (t) ≤ max max pj (t) ≤ b;
0≤t≤1 1≤j≤n 0≤t≤1 1≤j≤n

(H2) f : [0, 1] × R → R is a C function, f (t, x) = ◦(|x|) as x → 0;

n n 1

pro
(H3) there are constants µ > 0 such that a/2 − b/µ > 0 and r ≥ 0 such that
for |x| ≥ r,
0 < µF (t, x) ≤ xf (t, x).
(H4) f : [0, 1] × Rn → Rn is locally Lipschitz continuous.
We are to show in this paper the following results via variational method.
Theorem 1.1. Suppose assumptions (H1)-(H3) hold. BVPs (1.4) possesses a
nontrivial solution. re-
Theorem 1.2. Suppose assumptions (H1)-(H4) hold. BVPs (1.4) possesses a
positive and a negative solution.

2. Critical point of functional and solution of BVPs

Suppose X is a Banach space, Φ : X → R a differentiable functional with
derivative given by < Φ′ (u), v > with u, v ∈ X. If there is a u0 ∈ X such that
P
< Φ′ (u0 ), v > = 0 holds for all v ∈ X, then u0 is called a critical point of
functional Φ.
Let X = {x ∈ H 1 ([0, 1], Rn ) | x(0) = 0, x(1) = βx(η)} with the norm
Z 1 Z 1  12
al

2 ′ 2
kxk = |x(t)| dt + |x (t)| dt
0 0

for each x ∈ X. Then X is a closed subspace of reflexive Banach space

H 1 ([0, 1], Rn ), and so X itself is a reflexive Banach space.
rn

Lemma 2.1. [7] If x ∈ C 1 [0, 1] and x(0) = 0, then

Z 1 Z 1
4
|x(t)|2 dt ≤ |x′ (t)|2 dt. (2.1)
0 π2 0

Thanks to above Lemma, we can and will take as norm in X

Jou

Z 1  12
′ 2
kxk = |x (t)| dt .
0

Suppose U ⊂ Rn is a open subset. A function f : U → R is called a bump

function if it is both smooth(in the sense of having continuous derivatives of all
orders) and compactly supported (i.e., supp f ⊂ U ). The space of all bump
functions on U ⊂ Rn is denoted by C0∞ (U ).

3
 Journal Pre-proof

of
Lemma 2.2. [8](Paul du Bois-Reymend Lemma) Suppose f : U → R is a
locally integrable function on an open subset U ⊂ Rn , and suppose that
Z
f (t)φ(t)dt = 0
U

pro
for all φ ∈ C0∞ (U ). Then f (t) = 0 almost everywhere.
The following lemma is fundamental to the proof of our main theorem.
Lemma 2.3. If u ∈ X is a critical point of the functional
Z 1 
1 ′ ′
Φ(x) = (P (t)x (t), x (t)) − F (t, x(t)) dt, (2.2)
0 2
then u = u(t) is a solution of BVPs (1.4). re-
Proof. The properties of f and P ensure Φ is continuously differentiable and
the derivative of Φ is in the form
Z 1
< Φ′ (x), y > = [(P (t)x′ (t), y ′ (t)) − (f (t, x(t)), y(t))]dt, (2.3)
0

x, y ∈ X. Then the assumption that u is a critical point of Φ means that

< Φ′ (u), y >= 0, ∀ y ∈ X. (2.4)
P
Let T1 = (0, η), T2 = (η, 1) and
Zi = {z ∈ C ∞ ([0, 1]; Rn ) | supp z ⊂ Ti }, i = 1, 2.
Clearly, there is Zi ⊂ X. Relation (2.4) implies
al

< Φ′ (u), z >= 0, ∀ z ∈ Zi , i = 1, 2. (2.5)

Thus
Z
0 = [(P (t)u′ (t), z ′ (t)) − (f (t, u(t)), z(t))]dt
Ti
rn

Z
= [−((P (t)u′ (t))′ , z(t)) − (f (t, u(t)), z(t))]dt
Ti
Z
= − ((P (t)u′ (t))′ + f (t, u(t)), z(t))dt.
Ti

Since z ∈ Zi is arbitrary and

Jou

{z |Ti | z ∈ Zi } = C0∞ (Ti ; Rn )

which together with Lemma 2.2 deduce
(P (t)u′ (t))′ + f (t, u(t)) = 0, a.e. t ∈ Ti .
Take i = 1, 2, then it holds that
(P (t)u′ (t))′ + f (t, u(t)) = 0, a.e. t ∈ [0, 1]. (2.6)
Thus, u = u(t) is a solution to BVPs (1.4). The proof is complete. ¶
The task now is to discuss the existence of critical points of Φ in X. The
following estimation relation between two norms k · k∞ , k · k and injection
theorem are important to our results.

4
 Journal Pre-proof

of
Lemma 2.4. For each x ∈ X, it holds that
kxk∞ ≤ kxk. (2.7)
Proof. For each x = (x1 (t), x2 (t), · · · , xn (t)) ∈ X, there holds xi (0) = 0 and

pro
Z 1 Z 1  21
|xi (t)| ≤ |x′i (t)|dt ≤ |x′i (t)|2 dt , i = 1, 2, · · · , n,
0 0

thus kxk∞ ≤ kxk. The proof is complete. ¶

The following lemma is a variation version of Proposition 1.2 of Ref. [13].
Lemma 2.5. X ֒→֒→ C([0, 1], Rn ).
Proof. Given bounded set W ⊂ X, we will show that W is relatively compact
re-
in C([0, 1], Rn ).
Firstly, by Lemma 2.4, the injection of X into C([0, 1], Rn ), with its natural
norm k · k∞ , is continuous. So, W is bounded in C([0, 1], Rn ).
Secondly, X is a reflexive Banach space, which implies there exists a se-
quence {uk } ⊂ W, u ∈ X such that uk ⇀ u, in X. It follows that uk ⇀
u, in C([0, 1], Rn ). We claim that the sequence {uk } is equi-uniformly continu-
ous since for 0 ≤ s ≤ t ≤ 1, we have
P
Z t Z t  12
1
|uk (t) − uk (s)| ≤ |u′k (τ )|dτ ≤ (t − s) 2 |u′k (τ )|2 dτ
s s
1 1
≤ (t − s) 2 kuk k ≤ C(t − s) 2 .
By Ascoli-Arzela Theorem, {uk }, and W is relatively compact in C([0, 1], Rn ).
al

By the uniqueness of the weak limit in C([0, 1], Rn ), every uniformly convergent
subsequence of {uk } converges uniformly to u on [0, 1]. ¶
Theorem 2.1. [1](Mountain Pass Theorem) Let E be a real Banach space and
I ∈ C 1 (E, R) satisfying (PS)-condition. Suppose I(0) = 0 and
rn

(I1 ) there are constants ρ, α > 0 such that I |∂Bρ ≥ α, and

(I2 ) there is an e ∈ E \ Bρ such that I(e) ≤ 0.
Then I possesses a critical value c ≥ α. Moreover c can be characterized as
c = inf max I(u),
g∈Γ u∈g([0,1])
Jou

where Γ = {g ∈ C([0, 1], E) | g(0) = 0, g(1) = e}.

3. Proof of Theorems
Proof of Theorem 1.1.
Proof. We show that the functional Φ, defined in (2.2), satisfies (PS)-
condition, (I1 ) and (I2 ).
Firstly, to verify (I2 ), note by integrating condition (H3), there exist con-
stants c, L > 0, µ > 2 such that
F (t, x) ≥ c|x|µ − L, for t ∈ [0, 1], x ∈ Rn .

5
 Journal Pre-proof

of
So, Z Z
1 1
J(u) ≡ F (t, u(t))dt ≥ c |u(t)|µ dt − L, (3.1)
0 0

for all u ∈ X. Choosing any u ∈ X \ {0}, by (H1) and (3.1), one has

pro
Z 1 Z 1
bs2
Φ(su) ≤ |u′ (t)|2 dt − F (t, su(t))dt
2 0 0
Z 1
bs2
≤ kuk2 − csµ |u(t)|µ dt + L → −∞
2 0

as s → ∞. Hence (I2 ) holds.

Secondly, for (I1 ), by (H2), given ǫ = aπ 2 /8 > 0, there is a 0 < δ ≤ 1 such
that |u| ≤ δ implies
ǫ
re-
|F (t, u)| ≤ |u|2 .
2
Consequently, via Lemma 2.1
Z Z
ǫ 1 2ǫ 1 ′ 2
J(u) ≤ |u(t)|2 dt ≤ 2 |u (t)| dt,
2 0 π 0
P
so Z  
1
a 2ǫ a 2ǫ
Φ(u) ≥ kuk2 − 2 |u′ (t)|2 dt = − kuk2 .
2 π 0 2 π2
Taking into account Lemma 2.4, setting ρ = δ, α = aδ 2 /4, one has (I1 ) holds.
Thirdly, suppose {uk } ⊂ X is a sequence such that {Φ(uk )} is bounded (i.e.
al

|Φ(uk )| ≤ M ) and Φ′ (uk ) → 0 as k → ∞. We are to verify that {uk } possesses

a convergent subsequence in X.
Taking into account (2.2) and (2.3), one has
Z 1 
1 ′ ′
rn

Φ(u) = (P (t)u (t), u (t)) − F (t, u(t)) dt

0 2
Z 1
a 2
≥ kuk − F (t, u(t))dt; (3.2)
2 0

and
Jou

Z 1  
< Φ′ (u), u > = P (t)|u′ |2 − (f (t, u(t)), u(t)) dt
0
Z 1
≤ bkuk2 − (f (t, u(t)), u(t))dt. (3.3)
0

For k large, with u = uk , setting E = {t ∈ [0, 1] | |u(t)| ≥ r}, combing (3.2) and
(3.3) yields
M ≥ Φ(u)
  Z 1  
1 a b 1
≥ < Φ′ (u), u > + − kuk + 2
(f (t, u(t)), u(t)) − F (t, u(t)) dt
µ 2 µ 0 µ

6
 Journal Pre-proof

of
  Z  
1 ′ a b 2 1
≥ < Φ (u), u > + − kuk + (f (t, u(t)), u(t)) − F (t, u(t)) dt
µ 2 µ E µ
Z  
1
+ (f (t, u(t)), u(t)) − F (t, u(t)) dt. (3.4)
[0,1]\E µ

pro
By (H3), a/2 − b/µ > 0, the third term on the right in (3.4) is positive. The
fourth term is bounded by a constant independently of k. Hence (3.4) implies
{uk } is bounded in X. By Lemma 2.5, going if necessary, to a subsequence, we
assume that uk ⇀ u in X and uk → u in C([0, 1], Rn ). Then

< Φ′ (uk ) − Φ′ (u), uk − u >

=< Φ′ (uk ), uk − u > − < Φ′ (u), uk − u >→ 0, as k → ∞. (3.5)
Using (2.3) and assumption (H1), we have

< Φ′ (uk ) − Φ′ (u), uk − u >

Z 1
re-
= (P (t)(u′k (t) − u′ (t)), u′k (t) − u′ (t))dt
0
Z 1
− (f (t, uk (t)) − f (t, u(t)), uk (t) − u(t))dt
P
0
Z 1
≥a |u′k (t) − u′ (t)|2 dt
0
Z 1
− (f (t, uk (t)) − f (t, u(t)), uk (t) − u(t))dt. (3.6)
0
al

The fact that uk → u in C([0, 1], Rn ) implies

Z 1
(f (t, uk (t)) − f (t, u(t)), uk (t) − u(t))dt → 0,
rn

0
Z 1
|uk (t) − u(t)| → 0, for t = 0, η, 1; |uk (t) − u(t)|2 dt → 0,
0

as k → ∞. So by uk ∈ X and uk (0) = 0, uk (1) = βuk (η), we have

Jou

u(0) = 0, u(1) = βu(η).

On the other hand, from (3.5) and (3.6), we get
Z 1
|u′k (t) − u′ (t)|2 dt → 0, as k → ∞.
0

Therefore, it holds that uk → u in X. Then Φ satisfies (PS)-condition.

It is clear that Φ(0) = 0 and then Theorem 1.1 follows from Lemma 2.3 and
Theorem 2.1. ¶
Proof of Theorem 1.2.
Proof. Based upon above arguments, the proof of Theorem 1.2 is just
similar to the proof of Corollary 2.23 [14]. ¶

7
 Journal Pre-proof

of
References
[1] A. Ambrosetti and P. Rabinowitz, Dual variational methods in critical point
theory and applications, J. Funct. Anal. 14 (1973) 349–381.
[2] Z. Bai, Z. Du, S. Zhang, Iterative method for a class of fourth-order p-

pro
Laplacian beam equation, J. Appl. Anal. Comput. 9: 4 (2019) 1443–1453
[3] G. Bonanno, Multiple critical points theorems without the Palais-Smale
condition, J. Math. Anal. Appl. 299 (2004) 600–614.
[4] Y. Cui, Uniqueness of solution for boundary value problems for fractional
differential equations, Appl. Math. Lett., 51 (2016), 48–54.
[5] W. Feng and J. Webb, Solvability of a three-point nonlinear boundary value
problems at resonance, Nonlinear Anal.-Theor 30 (1997) 3227–3238.
re-
[6] J. Graef, S. Heidarkhani, L. Kong, A critical points approach to multiplicity
results for multipoint boundary value problems, Appl. Anal. 90(2011) 1909–
1925.
[7] G. Hardy, J. Littlewood, G. Polya, Inequalities, Cambridge University
Press, London (1967).
[8] L. Hömander, The analysis of linear partial differential operators, 2nd ed,
Springer-Verlag, 1990.
P
[9] H. Li and J. Sun, Positive solutions of superlinear semipositone nonlinear
boundary value problems, Comp. Math. Appl. 61 (2011), 2806–2815.
[10] X. Liu, M. Jia, W. Ge, The method of lower and upper solutions for mixed
fractional four-point boundary value problem with p-Laplacian operator,
Appl. Math. Lett., 65 (2017) 56–62.
al

[11] X. Liu, M. Jia, Solvability and numerical simulations for BVPs of fractional
coupled systems involving left and right fractional derivatives, Appl. Math.
Comput., 353 (2019) 230–242.
[12] V. Il’in and E. Moiseev, Nonlocal boundary value problem of the second
rn

kind for a Sturm-Liuville operator, Diff. Equat. 23 (1987) 979–987.

[13] J. Mawhin, M. Willem, Critical point theory and hamiltonian systems, Ap-
plied Mathematical Sciences, Vol 74, Springer, New York, 1989.
[14] P. Rabinowitz, Minimax methods in critical point theory with applications
to differential equations, CBMS Reg. Cof. Ser. in Math., 65, Amer. Math.
Jou

Soc., Provindence, 1986.

[15] Y. Tian, W. Ge, Multiple solutions for a second-order Sturm-Liouville
boundary value problem, Taiwanese J. Math. 11 (2007) 975–988.
[16] Y. Tian, W. Ge, Periodic solutions of non-autonomous second-order sys-
tems with a p-Laplacian, Nonlinear Anal.-Theor 66 (2007) 192–203.
[17] Y. Tian, W. Ge, Multiple solutions of Sturm-Liouville boundary value
problem via lower and upper solutions and variational methods, Nonlin-
ear Anal.-Theor 74 (2011) 6733–6746.
[18] Y. Wei, Q. Song, Z. Bai, Existence and iterative method for some fourth
order nonlinear boundary value problems, Appl. Math. Lett., 87 (2019),
101–107.

8
 Journal Pre-proof
*Credit Author Statement

Paper title: Existence of solution of a three-point boundary value problem

via variational approach

of
Author list: Wen Lian, Zhanbing Bai, Zengji Du

pro
Credit author statement

Wen Lian: Writing- Original draft preparation.

re-
Zhanbing Bai: Conceptualization, Methodology, Supervision.

Zengji Du: Investigation. Writing- Reviewing and Editing,

P
rn al
Jou