Mathematical and Computer Modelling 58 (2013) 1769–1776

Contents lists available at ScienceDirect

Mathematical and Computer Modelling

journal homepage: www.elsevier.com/locate/mcm

Solutions for a boundary value problem at resonance

on [0, ∞)
Junling Li a,∗ , Bingmei Liu a , Lishan Liu b
a
College of Sciences, China University of Mining and Technology, Xuzhou 221116, Jiangsu, China
b
School of Mathematical Sciences, Qufu Normal University, Qufu 273165, Shandong, China

article info abstract

Article history: An m-point boundary value problem with one-dimensional p-Laplacian at resonance on
Received 30 October 2011 [0, ∞) is considered. By establishing a continuation theorem and applying a suitable
Received in revised form 30 May 2013 homotopy, Leray–Schauder degree, a priori estimate, the existence of solutions to the above
Accepted 4 June 2013
problem is obtained.
© 2013 Elsevier Ltd. All rights reserved.
Keywords:
m-point
At resonance
Leray–Schauder degree
Continuation theorem

1. Introduction

Consider the following m-point boundary value problem (BVP) with one-dimensional p-Laplacian at resonance on the
half-line [0, ∞):

(φ(x (t ))) + f (t , x(t ), x (t )) = 0, t ∈ (0, ∞),

 ′ ′ ′

m − 2
 (1.1)
x (0) = ai x′ (ξi ), x′ (∞) = lim x′ (t ) = 0,
′

t →∞
i=1

where φ(s) = |s|p−2 s, p > 1, ai ∈ (0, 1] with i=1 ai = 1, 0 < ξ1 < ξ2 < · · · < ξm−2 ≤ 1 are given.
m−2
BVP (1.1) is at resonance in the sense that the associated m-point boundary value problem

(φ(x (t ))) = 0, t ∈ (0, ∞),

 ′ ′

m−
2

 x ′
(0) = ai x′ (ξi ), x′ (∞) = 0,
i=1

has a nontrivial solution x(t ) = ρ , where ρ ∈ R is an arbitrary constant.

Boundary value problems on [0, ∞) have received extensive attention, see [1–8] and references therein. On finite
intervals, nonlinear boundary value problems at resonance have received a great deal of attention, see [9–13] and their
references therein. However, boundary value problems at resonance on [0, ∞) are rarely studied and most works depend

∗ Corresponding author.
E-mail addresses: lijl2006@163.com (J.L. Li), lbm2009@cumt.edu.cn (B.M. Liu), lls@mial.qfnu.edu.cn (L.S. Liu).

0895-7177/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.mcm.2013.06.003
1770 J.L. Li, et al. / Mathematical and Computer Modelling 58 (2013) 1769–1776

on the coincidence degree theory [14,15]. Recently [7] set up the existence and uniqueness of unbounded solutions to the
following m-point boundary value problem at resonance on [0, ∞)

x (t ) + f (t , x(t ), x (t )) = 0, t ∈ (0, ∞),

 ′′ ′

m
 − 2

θ (x (0)) = ai θ (x′ (ξi )), x′ (∞) = lim x′ (t ) = 0,

′

t →∞
i=1

on the basis of homotopy, Leray–Schauder degree and a priori estimate, where θ : R → R is a strictly increasing and
continuous function and ai ∈ (0, 1] with i=1 ai = 1, 0 < ξ1 < ξ2 < · · · < ξm−2 ≤ 1 are given.
m−2
Inspired by the above papers, the existence of solutions to BVP (1.1) with one-dimensional p-Laplacian at resonance
on [0, ∞) is established. This paper generalizes [7] with the differentiable function x to BVP (1.1) with one-dimensional
p-Laplacian φ which may be non-smooth. In fact one-dimensional p-Laplacian φ brings new difficulties to a priori estimate
of solution to BVP (1.1) and to assumptions on f . Moreover, compared to the results in [9–16], our findings presented in this
paper has the following new features. Firstly, the nonlinear term f in BVP (1.1) may be superlinear with respect to the second
variable and the third variable. Secondly, our findings extend results of BVP at resonance on finite intervals to the case on
[0, ∞). Thirdly, this paper provides a continuation theorem to solve boundary value problems at resonance on [0, ∞), while
most papers are based on coincidence degree theory. Further, to overcome difficulties caused by resonance on [0, ∞), the
main techniques such as a priori estimate, homotopy and Leray–Schauder degree theory are used in this paper.

2. Preliminaries and lemmas

Throughout this paper, the nonlinear term f : [0, ∞) × R × R → R satisfies the following conditions:
(H1 ) for each (s, r ) ∈ R × R, the function f (·, s, r ) is measurable on [0, ∞);
(H2 ) for a.e. t ∈ [0, ∞), the function f (t , ·, ·) is continuous on R × R;
(H3 ) for each R > 0, there exist a Lebesgue integrable function ρR : [0, ∞) → [0, ∞) with ρR (t ) > 0, t ∈ (0, ∞) such that
|f (t , (1 + t )s, r )| ≤ ρR (t ),
for a.e. t ∈ [0, ∞) and all (s, r ) ∈ R × R with |s| ≤ R, |r | ≤ R.
Let
x(t )
 
1
C∞ [0, ∞) = x : [0, ∞) → R | x, x′ is continuous and lim = 0, lim x′ (t ) = 0 .
t →∞ 1+t t →∞

∀x ∈ C∞ [0, ∞), define ∥x∥∞ = max supt ∈[0,∞) |x(t )|/|1 + t |, supt ∈[0,∞) |x (t )| , and then C∞ [0, ∞) is a Banach space
1 1
  1 ′
1
with the norm ∥ · ∥∞ ; see [9].
To formulate our continuation theorem, for λ ∈ [0, 1], we consider the following BVP:

(φ(x′ (t )))′ + λf (t , x(t ), x′ (t )) = 0, t ∈ (0, ∞),


(2.1)
B(x, λ) = 0, x′ (∞) = 0,
1
where B : C∞ [0, ∞) × [0, 1] → R is an operator given by:
   
m−2
 m−2

B(x, λ) = λ x (0) − ′
ai x (ξi )
′
+ (1 − λ) φ(x (0)) −
′
ai φ(x (ξi )) .
′
(2.2)
i=1 i =1

We also define Ψ : C∞
1
[0, ∞) × [0, 1] → C∞
1
[0, ∞) by setting
  ∞  m−2
   ∞ 
Ψ (x, λ)(t ) = x(0) + φ −1
λ f (τ , x(τ ), x (τ ))dτ
′
− ai φ −1
λ f (τ , x(τ ), x (τ ))dτ
′
0 i=1 ξi

m−2
  ξi  t
  ∞ 
+ (1 − λ) ai f (τ , x(τ ), x′ (τ ))dτ + φ −1 λ f (τ , x(τ ), x′ (τ ))dτ ds, (2.3)
i =1 0 0 s

where φ −1 : R → R is the inverse mapping of φ . As in [7], it is easy to prove that Ψ : C∞

1
[0, ∞) × [0, 1] → C∞
1
[0, ∞) is
well defined.
We also define a function F : R → R as follows:
m−2
 ξi
F (ρ) = f (t , ρ, 0)dt , ∀ρ ∈ R. (2.4)
i=1 0
 J.L. Li, et al. / Mathematical and Computer Modelling 58 (2013) 1769–1776 1771

Lemma 2.1 ([17]). Let Ω is one bounded open set in a real Banach space E. If Ψ : Ω̄ × [0, 1] → E is continuous and for each
fixed λ ∈ [0, 1], Ψ (·, λ) : Ω̄ → E is a compact operator. Further, the continuity of Ψ at λ ∈ [0, 1] is uniform with respect to
x ∈ Ω̄ . Then Ψ : Ω̄ × [0, 1] → E is completely continuous.

1
Lemma 2.2 ([9]). Let D ⊂ C∞ [0, ∞). Then D is relatively compact in C∞
1
[0, ∞) if the following conditions hold:
1
(i) D is bounded in C∞ [0, ∞);
(ii) the functions belonging to {y | y(t ) = x(t )/(1 + t ), x ∈ D} and z | z (t ) = x′ (t ), x ∈ D are locally equicontinuous on
 
[0, ∞);
(iii) the functions from {y | y(t ) = x(t )/(1 + t ), x ∈ D} and z | z (t ) = x′ (t ), x ∈ D are equiconvergent at ∞, that is, for any
 
ε > 0 there exists T (ε) > 0 such that
|y(t ) − y(∞)| < ε, |z (t ) − z (∞)| < ε,
for all t > T (ε) and x ∈ D.

Lemma 2.3. The operator Ψ : C∞

1
[0, ∞) × [0, 1] → C∞
1
[0, ∞) is continuous.
Proof. The proof is similar to Lemma 2.3 in [7], so it is omitted here. 

Lemma 2.4. For any fixed λ ∈ [0, 1], Ψ (·, λ) : C∞

1
[0, ∞) → C∞
1
[0, ∞) is a compact operator.

Proof. According to the fact that φ, φ −1 are odd continuous and increasing from R to R, it is easy to prove this lemma by
the same way as Lemma 2.4 in [7]. So we omit the proof here. 

Lemma 2.5. For any bounded set Ω ⊂ C∞

1
[0, ∞), the operator Ψ : Ω̄ × [0, 1] → C∞
1
[0, ∞) is completely continuous.

Proof. For any bounded set Ω ⊂ C∞ 1

[0, ∞), for any x ∈ Ω , by Lemma 2.3, Ψ (x, ·) : [0, 1] → C∞ 1
[0, ∞) is continuous.
Moreover the continuity of Ψ at λ is uniform with respect to x ∈ Ω because f satisfies (H1 )–(H3 ). Then by Lemmas 2.3 and
2.4, it is easy to know that all conditions of Lemma 2.1 are satisfied. Therefore, Ψ : Ω̄ × [0, 1] → C∞1
[0, ∞) is completely
continuous. 

3. Main results

The Brouwer and Leray–Schauder degree shall be denoted by degB and degLS respectively.
Now we formulate the following continuation theorem for the solvability of the BVP (1.1).

Theorem 3.1. Let Ω ⊂ C∞

1
[0, ∞) be a bounded open set and set Ω0 = Ω ∩ R. Assume that the following three conditions are
satisfied:

(i) for each λ ∈ (0, 1), BVP (2.1) has no solution x ∈ ∂ Ω ;

(ii) the equation F (ρ) = 0 has no solution with ρ ∈ ∂ Ω ∩ R;
(iii) the Brouwer degree degB (F , Ω0 , 0) ̸= 0.

Then BVP (1.1) has at least one solution in Ω .

Proof. We divide our proof into three steps.

Step 1 In this part, we shall show that if x is a solution of BVP (2.1), then x satisfies x(t ) = Ψ (x, λ)(t ) for any t ∈ [0, ∞).
Suppose x be a solution to BVP (2.1) for some λ ∈ (0, 1]. Integrate the equation in (2.1) and use x′ (∞) = 0, then we
obtain that
 ∞
φ(x′ (t )) = λ f (τ , x(τ ), x′ (τ ))dτ . (3.1)
t
 ∞
By (3.1), x′ (t ) = φ −1 λ f (τ , x(τ ), x′ (τ ))dτ . Substituting x′ (t ) and φ(x′ (t )) into B(x, λ) = 0, we acquire that

t
 
  ∞  m−2
   ∞
λ φ −1
λ f (τ , x(τ ), x (τ ))dτ
′
− ai φ −1
λ f (τ , x(τ ), x (τ ))dτ
′
0 i =1 ξi
  
∞ m−2
  ∞
+ (1 − λ) λ f (τ , x(τ ), x′ (τ ))dτ − ai λ f (τ , x(τ ), x′ (τ ))dτ = 0.
0 i=1 ξi
1772 J.L. Li, et al. / Mathematical and Computer Modelling 58 (2013) 1769–1776

m−2
Since i =1 ai = 1, then
 ∞ m−2
  ∞ m−2
  ∞
λ f (τ , x(τ ), x (τ ))dτ −′
ai λ f (τ , x(τ ), x (τ ))dτ = λ
′
ai f (τ , x(τ ), x′ (τ ))dτ
0 i =1 ξi i=1 0

m−2
  ∞
− ai λ f (τ , x(τ ), x′ (τ ))dτ
i =1 ξi
m−2
  ξi
=λ ai f (τ , x(τ ), x′ (τ ))dτ .
i=1 0

Then B(x, λ) = 0 is equal to the following equality

  ∞  m−2
   ∞ 
φ −1 λ f (τ , x(τ ), x′ (τ ))dτ − ai φ − 1 λ f (τ , x(τ ), x′ (τ ))dτ
0 i =1 ξi

m−2
  ξi
+ (1 − λ) ai f (τ , x(τ ), x′ (τ ))dτ = 0. (3.2)
i=1 0
t  ∞ ∞
Substituting (3.2) into (2.3), then Ψ (x, λ) = x(0) + φ −1 λ s f (τ , x(τ ), x′ (τ ))dτ ds. According to (3.1), λ s f (τ ,

0
x(τ ), x′ (τ ))dτ = φ(x′ (s)) holds. Then
 t
Ψ (x, λ) = x(0) + φ −1 φ(x′ (s)) ds = x(t ).
 
0

As far, we have shown that if x is a solution of BVP (2.1), then x satisfies

x(t ) = Ψ (x, λ)(t ), t ∈ [0, ∞). (3.3)
Step 2 In this part, we shall show that if x ∈ C∞ [0, ∞) satisfies (3.3), then x is a solution of BVP (2.1).
1

Suppose that for some λ ∈ (0, 1], x ∈ C∞ 1

[0, ∞) satisfies (3.3). Let t = 0 in (2.3) and (3.3), we obtain that (3.2) is satisfied.
Next differentiate (2.3) and (3.3), then we get that (3.1) is satisfied. Let t → ∞ in (3.1), then we get that x′ (∞) = 0. Finally,
by taking t = 0 in (2.3) and (3.3), we obtain that x satisfies B(x, λ) = 0. Consequently x satisfies the boundary conditions in
BVP (2.1).
Since (3.1) implies that φ(x′ (t )) is absolutely continuous on [0, ∞), by differentiation, we obtain
(φ(x′ (t )))′ = −λf (t , x(t ), x′ (t )), t ∈ [0, ∞).
By far, we have proved that if x ∈ C∞ [0, ∞) satisfies (3.3), then x satisfies BVP (2.1).
1

Step 3 In this part, we shall show that BVP (1.1) has at least one solution in Ω .
If there is a function x ∈ ∂ Ω satisfying BVP (1.1), then Theorem 3.1 is completed. Therefore, let us assume that BVP (1.1)
has no solution on ∂ Ω . According to Step 1 and Step 2, for λ ∈ (0, 1], x is a solution of BVP (2.1) if and only if x(t ) ∈ C∞
1
[0, ∞)
is a solution of (3.3). By (i), we have
x ̸= Ψ (x, λ), for all x ∈ ∂ Ω and λ ∈ (0, 1).
According to the definition of B(x, λ) given in (2.2), if λ = 1, BVP (2.1) becomes BVP (1.1). BVP (1.1) has no solution on ∂ Ω ,
so
x ̸= Ψ (x, λ), for all x ∈ ∂ Ω and λ ∈ (0, 1].
Now we claim that x ̸= Ψ (x, 0) for all x ∈ ∂ Ω . Otherwise, there is x0 ∈ ∂ Ω such that x0 = Ψ (x0 , 0). Then by (2.3), we
have
m−1
  ξi
x0 (t ) = Ψ (x0 , 0)(t ) = x0 (0) + ai f (τ , x0 (τ ), x′0 (τ ))dτ . (3.4)
i=1 0

The above equality deduces that x′0 (t ) = 0 and x0 (t ) = ρ , where ρ ∈ ∂ Ω ∩ R is a constant. Then we obtain x0 (0) = ρ . Thus
(3.4) becomes the following equality
m−2
  ξi
ρ=ρ+ ai f (τ , ρ, 0)dτ = ρ + F (ρ). (3.5)
i=1 0

Therefore F (ρ) = 0, which contradicts assumption (ii). Consequently we obtain that

x ̸= Ψ (x, λ), for all x ∈ ∂ Ω and λ ∈ [0, 1].
 J.L. Li, et al. / Mathematical and Computer Modelling 58 (2013) 1769–1776 1773

By (3.4), if x ∈ Ω satisfies x(t ) − Ψ (x(t ), 0) = 0, then x = ρ and ρ ∈ Ω ∩ R = Ω0 . From the definition of F given
in (2.4) we know that ρ − Ψ (ρ, 0) = 0 is equal to F (ρ) = 0. By Lemma 2.5, Ψ : Ω × [0, 1] → C∞ 1
[0, ∞) is completely
continuous. Consequently, degLS (I − Ψ (·, 0), Ω , 0) = degB (I − Ψ (·, 0)|R , Ω0 , 0) = degB (F , Ω0 , 0). Finally it follows from
the homotopy invariance property of Leray–Schauder degree and assumption (iii) that
degLS (I − Ψ (·, 1), Ω , 0) = degLS (I − Ψ (·, 0), Ω , 0)
= degB (I − Ψ (·, 0)|R , Ω0 , 0)
= degB (F , Ω0 , 0)
̸= 0.
Therefore, the mapping Ψ1 ≡ Ψ (·, 1) : C∞ 1
[0, ∞) → C∞ 1
[0, ∞) has at least one fixed point in Ω . For λ = 1, BVP (2.1)
turns into BVP (1.1). Hence BVP (1.1) has at least one solution in Ω . Theorem 3.1 is finished. 

Theorem 3.2. Suppose that f : [0, ∞) × R × R → R satisfies the following conditions:

(A1 ) there exist nonnegative functions d1 (t ), d2 (t ), r (t ) with (1 + t )p−1 d1 (t ), d2 (t ), r (t ) ∈ L1 [0, ∞) such that
|f (t , u, v)| ≤ d1 (t )|u|p−1 + d2 (t )|v|p−1 + r (t ),
for a.e. t ∈ [0, ∞) and all u, v ∈ R;
(A2 ) there exist constants Γ > 0, A ≥ 0, B ≥ 0 and u0 > 0 such that
|f (t , u, v)| ≥ Γ |u|p−1 − A|v|p−1 − B,
for all u with |u| > u0 , all t ∈ [0, ∞) and all v ∈ R;
(A3 ) there exists a constant R > max {u0 , B/Γ } such that f (t , ρ, 0)f (t , −ρ, 0) < 0 for all ρ > R and all t ∈ [0, ∞);
(A4 ) Suppose that ∥d2 ∥L1 [0,∞) < 1 and the function Φ : [0, ∞) → [0, ∞) defined by
 ∞ 
(1 + t )p−1 d1 (t )dt B + Aφ(z )
 
∥r ∥L1 [0,∞)
Φ (z ) = φ −1 0
φ(z ) + + φ −1
1 − ∥d2 ∥L1 [0,∞) 1 − ∥d2 ∥L1 [0,∞) Γ
satisfies
lim sup Φ (z )/z < 1.
z →∞

1
Then BVP (1.1) has at least one solution x ∈ C∞ [0, ∞).
Proof. We shall show that under assumptions (A1 )–(A4 ), there exists a bounded open set Ω ⊂ C∞ 1
[0, ∞) and Ω satisfies
all conditions of Theorem 3.1.
1
Now we need to make a priori estimate for the solutions of BVP (2.1). Let x ∈ C∞ [0, ∞) be a solution to BVP (2.1) for some
λ ∈ (0, 1]. Then integrate the equation in (2.1) from t to ∞, use x (∞) = 0, (A1 ) and the fact that φ is odd and increasing
′

from R to R, then we obtain that for any t ∈ [0, ∞)

 ∞
φ(|x′ (t )|) ≤ d1 (t )|x(t )|p−1 + d2 (t )|x′ (t )|p−1 + r (t ) dt
 

0 ∞
(1 + t )p−1 d1 (t )(∥x∥1∞ )p−1 + d2 (t )|x′ (t )|p−1 dt + ∥r ∥L1 [0,∞)
 
≤
0
 ∞  
≤ (1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥d2 ∥L1 [0,∞) φ max |x′ (t )| + ∥r ∥L1 [0,∞) .
0 t ∈[0,∞)

That deduces
   ∞  
φ max |x′ (t )| ≤ (1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥d2 ∥L1 [0,∞) φ max |x′ (t )| + ∥r ∥L1 [0,∞) .
t ∈[0,∞) 0 t ∈[0,∞)

Thus
∞
(1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥r ∥L1 [0,∞)
 
φ max |x′ (t )| ≤ 0
.
t ∈[0,∞) 1 − ∥d2 ∥L1 [0,∞)

Since φ −1 is increasing from R to R, we obtain that

 ∞ 
(1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥r ∥L1 [0,∞)
max |x′ (t )| ≤ φ −1 0
. (3.6)
t ∈[0,∞) 1 − ∥d2 ∥L1 [0,∞)
1774 J.L. Li, et al. / Mathematical and Computer Modelling 58 (2013) 1769–1776

From (A1 ) it follows that for x ∈ C∞

1
[0, ∞) and any t ∈ [0, ∞)
 ∞  ∞ 
f (τ , x(τ ), x (τ ))dτ ≤ ′
(1 + τ ) p−1
d1 (τ )dτ + ∥d2 ∥L1 [0,∞) φ(∥x∥1∞ ) + ∥r ∥L1 [0,∞) < ∞.
t 0

Integrate the Eq. (2.1) from t to ∞ and use x′ (∞) = 0, then we obtain that
 ∞
φ(x′ (t )) = λ f (τ , x(τ ), x′ (τ ))dτ .
t

Then B(x, λ) = 0 becomes

  ∞  m−2
   ∞ 
φ −1 λ f (τ , x(τ ), x′ (τ ))dτ − ai φ − 1 λ f (τ , x(τ ), x′ (τ ))dτ
0 i =1 ξi

m−2
  ξi
+ (1 − λ) ai f (τ , x(τ ), x′ (τ ))dτ = 0. (3.7)
i=1 0

We assert that there exists a τ0 ∈ [0, ∞) such that f (τ0 , x(τ0 ), x′ (τ0 )) = 0. In fact, if it is not true, without loss of
generality, we assume that f (τ , x(τ ), x′ (τ )) > 0 for all τ ∈ [0, ∞). Set
m−2
   ∞ 
Iλ = ai φ − 1 λ f (τ , x(τ ), x′ (τ ))dτ .
i =1 ξi

Since f (τ , x(τ ), x′ (τ )) > 0 for all τ ∈ [0, ∞), the following two equalities hold:
 ∞  ∞ 
f (τ , x(τ ), x (τ ))dτ = min ′
f (τ , x(τ ), x (τ ))dτ : i = 1, 2, . . . , m − 2
′
ξm−2 ξi

and
 ∞  ∞ 
f (τ , x(τ ), x (τ ))dτ = max
′
f (τ , x(τ ), x (τ ))dτ : i = 1, 2, . . . , m − 2 .
′
ξ1 ξi
m−2
Then by i =1 ai = 1 and the expression of Iλ , we obtain that
  
∞   ∞ 
φ −1
λ f (τ , x(τ ), x (τ ))dτ ′
≤ Iλ ≤ φ −1
λ f (τ , x(τ ), x (τ ))dτ
′
.
ξm−2 ξ1
 ∞
Let S (t ) = φ −1 λ t f (τ , x(τ ), x′ (τ ))dτ . Then S (t ) is a continuous nonincreasing function and we obtain that S (ξm−2 ) ≤

Iλ ≤ S (ξ1 ). Then by Intermediate Value Theorem, there exists tλ ∈ [ξ1 , ξm−2 ] such that Iλ = S (tλ ), i.e.
m−2
   ∞    ∞ 
ai φ − 1 λ f (τ , x(τ ), x′ (τ ))dτ = φ −1 λ f (τ , x(τ ), x′ (τ ))dτ .
i=1 ξi tλ

Similarly there is η0 ∈ [ξ1 , ξm−2 ] such that

m−2
  ξi  η0
ai f (τ , x(τ ), x (τ ))dτ = ′
f (τ , x(τ ), x′ (τ ))dτ .
i =1 0 0

By the above two equalities, for λ ∈ (0, 1], (3.7) can be rewritten as
  ∞    ∞   η0
φ −1
λ f (τ , x(τ ), x (τ ))dτ ′
−φ −1
λ f (τ , x(τ ), x (τ ))dτ ′
+ (1 − λ) f (τ , x(τ ), x′ (τ ))dτ = 0. (3.8)
0 tλ 0
 η0
Since (1 − λ) 0
f (τ , x(τ ), x′ (τ ))dτ ≥ 0 for λ ∈ (0, 1], by (3.8), we obtain that
  ∞    ∞ 
φ −1
λ f (τ , x(τ ), x (τ ))dτ ′
≤φ −1
λ f (τ , x(τ ), x (τ ))dτ ′
.
0 tλ

Because φ is increasing and f (τ , x(τ ), x′ (τ )) > 0 for all τ ∈ [0, ∞), we derive that
 ∞  ∞  ∞
f (τ , x(τ ), x′ (τ ))dτ ≤ f (τ , x(τ ), x′ (τ ))dτ < f (τ , x(τ ), x′ (τ ))dτ ,
0 tλ 0

which is a contradiction. So there exists τ0 ∈ [0, ∞) such that f (τ0 , x(τ0 ), x′ (τ0 )) = 0.
 J.L. Li, et al. / Mathematical and Computer Modelling 58 (2013) 1769–1776 1775

Take τ1 = min τ ∈ [0, ∞) | f (τ , x(τ ), x′ (τ )) = 0 . Obviously τ1 ≤ τ0 .

 
Replacing θ with φ in [7], we can prove τ1 ≤ ξm−2 easily.
In what follows, we estimate ∥x∥1∞ in two cases: |x(τ1 )| > u0 and |x(τ1 )| ≤ u0 .
If |x(τ1 )| > u0 , then by (A2 ), we obtain that

B + Aφ(|x′ (τ1 )|)

φ(|x(τ1 )|) ≤ . (3.9)
Γ
For any t ∈ [0, ∞),
 t
x(t ) = x(τ1 ) + x′ (s)ds.
τ1

Therefore, by (3.6) and (3.9), for any t ≥ τ1

 x(t ) 
 ≤ sup |x(τ1 )| + sup
   t
1
sup  |x′ (s)|ds ≤ |x(τ1 )| + max |x′ (s)|,
t ∈[0,∞) 1 + t s∈[0,∞)
 t ∈[0,∞) 1 + t t ∈[0,∞) 1 + t τ1

and for any t < τ1

 x(t )   τ1 − t 
   
 ≤ |x(τ1 )| + t ∈[sup
sup     max |x′ (s)| ≤ |x(τ1 )| + τ1 max |x′ (s)|.
t ∈[0,∞) 1 + t 0,∞) 1 + t s∈[0,∞) s∈[0,∞)
 

According to the above two inequalities, (3.6), (3.9) and τ1 ≤ ξm−2 ≤ 1, we obtain that
  
B + Aφ max |x′ (t )|
 x(t ) 
 
t ∈[0,∞)
 ≤ φ −1 
 
sup  
t ∈[0,∞) 1 + t
  Γ 

 ∞ 
(1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥r ∥L1 [0,∞)
+ max{1, τ1 }φ −1 0
1 − ∥d2 ∥L1 [0,∞)
B + Aφ( max |x (t )|)
′
 
t ∈[0,∞)
≤ φ −1  
Γ
 ∞ 
(1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥r ∥L1 [0,∞)
+φ −1 0
. (3.10)
1 − ∥d2 ∥L1 [0,∞)

The definition of ∥ · ∥1∞ together with (3.6) and (3.10) shows that
 ∞ 
B + Aφ(∥x∥1∞ ) (1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥r ∥L1 [0,∞)
 
∥x ∥∞ ≤ φ
1 −1
+φ −1 0
Γ 1 − ∥d2 ∥L1 [0,∞)

= Φ (∥x∥1∞ ). (3.11)

If |x(τ1 )| ≤ u0 , like in the proof of the case |x(τ1 )| > u0 , it is easy to prove that
 x(t ) 
 
sup   ≤ u0 + max |x′ (t )|.
t ∈[0,∞) 1 + t t ∈[0,∞)


Then by (3.6), we obtain

 ∞ 
(1 + t )p−1 d1 (t )dt · φ(∥x∥1∞ ) + ∥r ∥L1 [0,∞)
∥ x ∥ ∞ ≤ u0 + φ
1 −1 0
. (3.12)
1 − ∥d2 ∥L1 [0,∞)

It follows from (3.11) and (3.12) that in any case, ∥x∥1∞ satisfies the following inequality

∥x∥1∞ ≤ u0 + Φ (∥x∥1∞ ). (3.13)

By (A4 ) : lim supz →∞ Φ (z )/z < 1 and (3.13), there exists a constant C > 0 independent of λ ∈ (0, 1] such that
∥x∥1∞ ≤ C .
1776 J.L. Li, et al. / Mathematical and Computer Modelling 58 (2013) 1769–1776

Up to now, it has been shown that for λ ∈ (0, 1], if x ∈ C∞

1
[0, ∞) is a solution of BVP (2.1), then
∥x∥1∞ ≤ C ,
where C is a constant independent of λ ∈ (0, 1].
Taking R0 > max{R, C } and Ω = B(0, R0 ) ⊂ C∞ 1
[0, ∞), then for λ ∈ (0, 1], BVP (2.1) has no solution on ∂ Ω , where R is
given in (A3 ). Accordingly, the condition (i) of Theorem 3.1 is satisfied.
We see from assumption (A3 ) that |f (t , ρ, 0)| > 0 for all |ρ| > R and all t ∈ [0, ∞). Therefore, by (2.4),
m−2
  ξi
F (ρ) = ai f (t , ρ, 0)dt (3.14)
i=1 0

is either strictly positive or strictly negative for all ρ with |ρ| > R. Consequently, the equation F (ρ) = 0 has no solution for
ρ ∈ ∂ Ω ∩ R. This means that the condition (ii) of Theorem 3.1 is satisfied.
Note that R0 > R. Then by (A3 ) and (3.14), F (R0 ) and F (−R0 ) have opposite signs. Therefore, Brouwer degree
degB (F , Ω0 , 0) ̸= 0 holds, where Ω0 = Ω ∩ R. That is, the condition (iii) of Theorem 3.1 is satisfied.
So far, all conditions of Theorem 3.1 have been satisfied. Consequently, by Theorem 3.1, BVP (1.1) has at least one solution
in Ω . 

Remark 3.1. The condition (A1 ) shows that the nonlinear term f (t , u, v) in BVP (1.1) may be superlinear with respect to
u, v and solutions obtained by Theorem 3.1 may be unbounded.

Theorem 3.3. Suppose that f : [0, ∞) × R × R → R satisfies (A3 ), (A4 ) and the following two conditions:
(A5 ) there exist nonnegative functions d3 (t ), d4 (t ) with (1 + t )p−1 d3 (t ), d4 (t ) ∈ L1 [0, ∞) such that
|f (t , u1 , v1 ) − f (t , u2 , v2 )| ≤ d3 (t )|u1 − u2 |p−1 + d4 (t )|v1 − v2 |p−1 ,
for a.e. t ∈ [0, ∞) and all u1 , u2 , v1 , v2 ∈ R;
(A6 ) there exist constants Γ1 > 0, A1 ≥ 0 and u0 > 0 such that
|f (t , u1 , v1 ) − f (t , u2 , v2 )| ≥ Γ1 φ(|u1 − u2 |) − A1 φ(|v1 − v2 |),
for all u1 , u2 with |u1 − u2 | > u0 , all t ∈ [0, ∞) and all v1 , v2 ∈ R.
Then BVP (1.1) has at least one solution x ∈ C∞ 1
[0, ∞).
It is straightforward to prove the Theorem 3.3, so we omit the proof here.

Acknowledgments

The authors are supported financially by the National Natural Science Foundation of China (11201473, 11071141,
11271364) and the Fundamental Research Funds for the Central Universities (2013QNA35, 2010LKSX09).

References

[1] R.P. Agarwal, D. O’Regan, Infinite Interval Problems for Differential Difference and Integral Equations, Kluwer Academic, 2001.
[2] S.Z. Chen, Y. Zhang, Singular boundary value problem on a half-line, J. Math. Anal. Appl. 195 (1995) 449–468.
[3] L.E. Bobisud, Existence of positive solutions to some nonlinear singular boundary value problems on finite and infinite intervals, J. Math. Anal. Appl.
173 (1993) 69–83.
[4] J.V. Baxley, Existence and uniqueness of nonlinear boundary value problems on infinite intervals, J. Math. Anal. Appl. 147 (1990) 127–133.
[5] B.M. Liu, L.S. Liu, Y.H. Wu, Multiple solutions of singular three-point boundary value problems on [0, ∞), Nonlinear Anal. 70 (2009) 3348–3357.
[6] B.M. Liu, L.S. Liu, Y.H. Wu, Unbounded solutions for three-point boundary value problems with nonlinear conditions on [0, +∞), Nonlinear Anal. 73
(2009) 2923–2932.
[7] B.M. Liu, L.S. Liu, Y.H. Wu, Existence and uniqueness for an m-point boundary value problem at resonance on infinite intervals, Comput. Math. Appl.
64 (2012) 1677–1690.
[8] D. O’Regan, B.Q. Yan, R.P. Agarwal, Solutions in weighted spaces of singular boundary value problems on the half-line, J. Comput. Appl. Math. 205
(2007) 751–763.
[9] B.Q. Yan, D. O’Regan, R.P. Agarwal, Unbounded solutions for singular boundary value problems on the semi-infinite interval: upper and lower solutions
and multiplicity, J. Comput. Appl. Math. 197 (2006) 365–386.
[10] C.Z. Bai, J.X. Fang, Existence of positive solutions for three-point boundary value problems at resonance, J. Math. Anal. Appl. 291 (2004) 538–549.
[11] Z.J. Du, X.J. Lin, W.G. Ge, On a third-order multi-point boundary value problem at resonance, J. Math. Anal. Appl. 302 (2005) 217–229.
[12] M. García-Huidobro, C.P. Gupta, R. Manásevich, An m-point boundary value problem of Neumann type for a p-Laplacian like operator, Nonlinear Anal.
56 (2004) 1071–1089.
[13] B. Liu, Solvability of multi-point boundary value problem at resonance (III), Appl. Math. Comput. 129 (2002) 119–143.
[14] R.Y. Ma, Existence results of a m-point boundary value problem at resonance, J. Math. Anal. Appl. 294 (2004) 147–157.
[15] H.R. Lian, H.H. Pang, W.G. Ge, Solvability for second-order three-point boundary value problems at resonance on a half-line, J. Math. Anal. Appl. 337
(2008) 1171–1181.
[16] Nickolai Kosmatov, Multi-point boundary value problems on an unbounded domain at resonance, Nonlinear Anal. 68 (2008) 2158–2171.
[17] D.J. Guo, Nonlinear Functional Analysis, second ed., Shandong Science and Technology Press, Jinan, 2003 (in Chinese).