Abstract

We obtain new result of the existence of positive solutions of a class of singular impulse periodic boundary value problem via a nonlinear alternative principle of Leray-Schauder. We do not require the monotonicity of functions in paper (Zhang and Wang, 2003). An example is also given to illustrate our result.

1. Introduction

Because of wide interests in physics and engineering, periodic boundary value problems have been investigated by many authors (see [1โ€“19]). In most real problems, only the positive solution is significant.

In this paper, we consider the following periodic boundary value problem (PBVP in short) with impulse effects: โˆ’๐‘ข๎…ž๎…ž||(๐‘ก)+๐‘€๐‘ข(๐‘ก)=๐‘“(๐‘ก,๐‘ข(๐‘ก)),๐‘กโˆˆ๐ฝโ€ฒ,ฮ”๐‘ข๐‘ก=๐‘ก๐‘˜=๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ,โˆ’ฮ”๐‘ข๎…ž||๐‘ก=๐‘ก๐‘˜=๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ,๐‘˜=1,2,โ€ฆ๐‘™,๐‘ข(0)=๐‘ข(2๐œ‹),๐‘ข๎…ž(0)=๐‘ข๎…ž(2๐œ‹).(1.1) Here, ๐ฝ=[0,2๐œ‹], 0<๐‘ก1<๐‘ก2<โ‹ฏ<๐‘ก๐‘™<2๐œ‹, ๐ฝโ€ฒ=๐ฝโงต{๐‘ก1,๐‘ก2,โ€ฆ,๐‘ก๐‘™}, ๐‘€>0, ๐‘“โˆˆ๐ถ(๐ฝร—๐‘…+,๐‘…+), ๐ผ๐‘˜โˆˆ๐ถ(๐‘…+,๐‘…), ๐ฝ๐‘˜โˆˆ๐ถ(๐‘…+,๐‘…+), ๐‘…+=[0,+โˆž), ๐‘…+=(0,+โˆž) with โˆ’(1/๐‘š)๐ฝ๐‘˜(๐‘ข)<๐ผ๐‘˜(๐‘ข)<(1/๐‘š)๐ฝ๐‘˜(๐‘ข), ๐‘ขโˆˆ๐‘…+, โˆš๐‘š=๐‘€. ฮ”๐‘ข|๐‘ก=๐‘ก๐‘˜=๐‘ข(๐‘ก+๐‘˜)โˆ’๐‘ข(๐‘กโˆ’๐‘˜), ฮ”๐‘ข๎…ž|๐‘ก=๐‘ก๐‘˜=๐‘ขโ€ฒ(๐‘ก+๐‘˜)โˆ’๐‘ขโ€ฒ(๐‘กโˆ’๐‘˜), where ๐‘ข(๐‘–)(๐‘ก+๐‘˜) and ๐‘ข(๐‘–)(๐‘กโˆ’๐‘˜), ๐‘–=0,1, respectively, denote the right and left limit of ๐‘ข(๐‘–)(๐‘ก) at ๐‘ก=๐‘ก๐‘˜.

In [7], Liu applied Krasnoselskii's and Leggett-Williams fixed-point theorem to establish the existence of at least one, two, or three positive solutions to the first-order periodic boundary value problems๐‘ฅ๎…ž[]โงต๎€ฝ๐‘ก(๐‘ก)+๐‘Ž(๐‘ก)๐‘ฅ(๐‘ก)=๐‘“(๐‘ก,๐‘ฅ(๐‘ก)),a.e.๐‘กโˆˆ0,๐‘‡1,โ€ฆ,๐‘ก๐‘๎€พ,||ฮ”๐‘ฅ๐‘ก=๐‘ก๐‘˜=๐ผ๐‘˜๎€ท๐‘ฅ๎€ท๐‘ก๐‘˜๐‘ฅ๎€ธ๎€ธ,๐‘˜=1,โ€ฆ,๐‘,(0)=๐‘ฅ(๐‘‡).(1.2) Jiang [5] has applied Krasnoselskii's fixed point theorem to establish the existence of positive solutions of problem ๐‘ฅ๎…ž๎…ž[],(๐‘ก)+๐‘€๐‘ฅ(๐‘ก)=๐‘“(๐‘ก,๐‘ฅ(๐‘ก)),๐‘กโˆˆ0,2๐œ‹๐‘ฅ(0)=๐‘ฅ(2๐œ‹),๐‘ฅ๎…ž(0)=๐‘ฅ๎…ž(2๐œ‹).(1.3) The work [5] proved that periodic boundary value problem (PBVP in short) (1.3) without singularity have at least one positive solutions provided ๐‘“(๐‘ก,๐‘ฅ) is superlinear or sublinear at ๐‘ฅ=0+ and ๐‘ฅ=+โˆž. In [14], Tian et al. researched PBVP (1.1) without singularity. They obtained the existence of multiple positive solutions of PBVP (1.1) by replacing the suplinear condition or sublinear condition of [4] with the following limit inequality condition:(๐ด1)๎ƒฌ2๐œ‹๐‘“0+๐‘™๎“๐‘–=1๐ฝ0๎ƒญ๎ƒฌ(๐‘–)๐œŽ>2๐œ‹๐‘€,2๐œ‹๐‘“โˆž+๐‘™๎“๐‘–=1๐ฝโˆž๎ƒญ(๐‘–)๐œŽ>2๐œ‹๐‘€,(1.4)(๐ด2)๎ƒฌ2๐œ‹๐‘“0+๐‘™๎“๐‘–=1๐ฝ0๎ƒญ๎ƒฌ(๐‘–)๐œŽ<2๐œ‹๐‘€,2๐œ‹๐‘“โˆž+๐‘™๎“๐‘–=1๐ฝโˆž๎ƒญ(๐‘–)๐œŽ<2๐œ‹๐‘€.(1.5)

Nieto [10] introduced the concept of a weak solution for a damped linear equation with Dirichlet boundary conditions and impulses. These results will allow us in the future to deal with the corresponding nonlinear problems and look for solutions as critical points of weakly lower semicontinuous functionals.

We note that the function ๐‘“ involved in above papers [5, 7, 10, 14] does not have singularity. Xiao et al. [16] investigate the multiple positive solutions of singular boundary value problem for second-order impulsive singular differential equations on the halfline, where the function ๐‘“(๐‘ก,๐‘ข) is singular only at ๐‘ก=0 and/or ๐‘ก=1. Reference [19] studied PBVP (1.3), where the function ๐‘“ has singularity at ๐‘ฅ=0. The authors present the existence of multiple positive solutions via the Krasnoselskii's fixed point theorem under the following conditions.(๐ด๎…ž1)There exist nonnegative valued ๐œ‰(๐‘ฅ),๐œ‚(๐‘ฅ)โˆˆ๐ถ((0,โˆž)) and ๐‘ƒ(๐‘ก),๐‘„(๐‘ก)โˆˆ๐ฟ1[0,2๐œ‹] such that []0โ‰ค๐‘“(๐‘ก,๐‘ฅ)โ‰ค๐‘ƒ(๐‘ก)๐œ‰(๐‘ฅ)+๐‘„(๐‘ก)๐œ‚(๐‘ฅ),a.e.(๐‘ก,๐‘ฅ)โˆˆ0,2๐œ‹ร—(0,โˆž),sup๐‘ฅโˆˆ(0,โˆž)โŽงโŽชโŽจโŽชโŽฉ๐‘ฅ๎‚€โˆซ02๐œ‹โˆซ๐‘ƒ(๐‘ก)๐‘‘๐‘ก๐œ‰(๐‘ฅ)/๐œ‚(๐‘ฅ)+02๐œ‹๎‚๐œ‚๎€ท๐›ฟ๐‘„(๐‘ก)๐‘‘๐‘ก๐‘—๐‘ก๎€ธโŽซโŽชโŽฌโŽชโŽญ>๐ต๐‘—,(1.6)

where ๐œ‚(๐‘ฅ) is nonincreasing and ๐œ‰(๐‘ฅ)/๐œ‚(๐‘ฅ) is nondecreasing on (0,โˆž),(๐ด๎…ž2)lim๐‘กโ†’0+๎‚†โˆซinfmin02๐œ‹๐‘“(๐‘ฅ,๐‘ค)๐‘‘๐‘ฅโˆถ๐›ฟ๐‘—๎‚‡๐‘กโ‰ค๐‘คโ‰ค๐‘ก๐‘ก>1๐ด๐‘—,(1.7)(๐ด๎…ž3)lim๐‘กโ†’+โˆž๎‚†โˆซinfmin02๐œ‹๐‘“(๐‘ฅ,๐‘ค)๐‘‘๐‘ฅโˆถ๐›ฟ๐‘—๎‚‡๐‘กโ‰ค๐‘คโ‰ค๐‘ก๐‘ก>1๐ด๐‘—.(1.8) Here, ๐›ฟ๐‘—,๐ด๐‘—,๐ต๐‘— are some constants.

In this paper, the nonlinear term ๐‘“(๐‘ก,๐‘ข) is singular at ๐‘ข=0, and positive solution of PBVP (1.1) is obtained by a nonlinear alternative principle of Leray-Schauder type in cone. We do not require the monotonicity of functions ๐œ‚, ๐œ‰/๐œ‚ used in [19]. An example is also given to illustrate our result.

This paper is organized as follows. In Section 1, we give a brief overview of recent results on impulsive and periodic boundary value problems. In Section 2, we present some preliminaries such as definitions and lemmas. In Section 3, the existence of one positive solution for PBVP (1.1) will be established by using a nonlinear alternative principle of Leray-Schauder type in cone. An example is given in Section 4.

2. Preliminaries

Consider the space ๐‘ƒ๐ถ[๐ฝ,๐‘…]={๐‘ขโˆถ๐‘ข is a map from ๐ฝ into ๐‘… such that ๐‘ข(๐‘ก) is continuous at ๐‘กโ‰ ๐‘ก๐‘˜, left continuous at ๐‘ก=๐‘ก๐‘˜, and ๐‘ข(๐‘ก+๐‘˜) exists, for ๐‘˜=1,2,โ€ฆ๐‘™.}. It is easy to say that ๐‘ƒ๐ถ[๐ฝ,๐‘…] is a Banach space with the norm โ€–๐‘ขโ€–๐‘๐‘=sup๐‘กโˆˆ๐ฝ|๐‘ข(๐‘ก)|. Let ๐‘ƒ๐ถ1[๐ฝ,๐‘…]={๐‘ขโˆˆ๐‘ƒ๐ถ[๐ฝ,๐‘…]โˆถ๐‘ขโ€ฒ(๐‘ก) exists at ๐‘กโ‰ ๐‘ก๐‘˜ and is continuous at ๐‘กโ‰ ๐‘ก๐‘˜, and ๐‘ข๎…ž(๐‘ก+๐‘˜), ๐‘ข๎…ž(๐‘กโˆ’๐‘˜) exist and ๐‘ข๎…ž(๐‘ก) is left continuous at ๐‘ก=๐‘ก๐‘˜, for ๐‘˜=1,2,โ€ฆ๐‘™.} with the norm โ€–๐‘ขโ€–๐‘๐‘1=max{โ€–๐‘ขโ€–๐‘๐‘,โ€–๐‘ขโ€ฒโ€–๐‘๐‘}. Then, ๐‘ƒ๐ถ1[๐ฝ,๐‘…] is also a Banach space.

Lemma 2.1 (see [15]). ๐‘ขโˆˆ๐‘ƒ๐ถ1(๐ฝ,๐‘…)โˆฉ๐ถ2(๐ฝโ€ฒ,๐‘…) is a solution of PBVP (1.1) if and only if ๐‘ขโˆˆ๐‘ƒ๐ถ(๐ฝ) is a fixed point of the following operator ๐‘‡: ๎€œ๐‘‡๐‘ข(๐‘ก)=02๐œ‹๐บ(๐‘ก,๐‘ )๐‘“(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ +๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜,๎€ธ๎€ธ(2.1) where ๐บ(๐‘ก,๐‘ ) is the Green's function to the following periodic boundary value problem: โˆ’๐‘ข๎…ž๎…ž+๐‘€๐‘ข=0,๐‘ข(0)=๐‘ข(2๐œ‹),๐‘ข๎…ž(0)=๐‘ข๎…ž1(2๐œ‹),๐บ(๐‘ก,๐‘ )โˆถ=ฮ“โŽงโŽชโŽจโŽชโŽฉโŽงโŽชโŽจโŽชโŽฉ๐‘’๐‘š(๐‘กโˆ’๐‘ )+๐‘’๐‘š(2๐œ‹โˆ’๐‘ก+๐‘ )๐‘’,0โ‰ค๐‘ โ‰ค๐‘กโ‰ค2๐œ‹,๐‘š(๐‘ โˆ’๐‘ก)+๐‘’๐‘š(2๐œ‹โˆ’๐‘ +๐‘ก),0โ‰ค๐‘กโ‰ค๐‘ โ‰ค2๐œ‹,(2.2) here, ฮ“=2๐‘š(๐‘’2๐‘š๐œ‹โˆ’1). It is clear that 2๐‘’๐‘š๐œ‹ฮ“๐‘’=๐บ(๐œ‹)โ‰ค๐บ(๐‘ก,๐‘ )โ‰ค๐บ(0)=2๐‘š๐œ‹+1ฮ“.(2.3) Define ๎€ฝ[]๐พ=๐‘ขโˆˆ๐‘ƒ๐ถ๐ฝ,๐‘…โˆถ๐‘ข(๐‘ก)โ‰ฅ๐œŽโ€–๐‘ขโ€–๐‘๐‘๎€พ,๐‘กโˆˆ๐ฝ,(2.4) where 1๐œŽ=๐‘’2๐‘š๐œ‹.(2.5)

The following nonlinear alternative principle of Leray-Schauder type in cone is very important for us.

Lemma 2.2 (see [4]). Assume that ฮฉ is a relatively open subset of a convex set ๐พ in a Banach space ๐‘ƒ๐ถ[๐ฝ,๐‘…]. Let ๐‘‡โˆถฮฉโ†’๐พ be a compact map with 0โˆˆฮฉ. Then, either (i)๐‘‡ has a fixed point in ฮฉ, or,(ii)there is a ๐‘ขโˆˆ๐œ•ฮฉ and a ๐œ†<1 such that ๐‘ข=๐œ†๐‘‡๐‘ข.

3. Main Results

In this section, we establish the existence of positive solutions of PBVP (1.1).

Theorem 3.1. Assume that the following three hypothesis hold: (๐ป1)there exists nonnegative functions ๐œ‰(๐‘ข),๐œ‚(๐‘ข),๐›พ(๐‘ข)โˆˆ๐ถ(0,+โˆž) and ๐‘(๐‘ก),๐‘ž(๐‘ก)โˆˆ๐ฟ1([0,2๐œ‹]) such that[]๐‘“(๐‘ก,๐‘ข)โ‰ค๐‘(๐‘ก)๐œ‰(๐‘ข)+๐‘ž(๐‘ก)๐œ‚(๐‘ข),(๐‘ก,๐‘ข)โˆˆ0,2๐œ‹ร—(0,โˆž),(3.1)max1โ‰ค๐‘˜โ‰ค๐‘™๐ฝ๐‘˜[](๐‘ข)โ‰ค๐›พ(๐‘ข),(๐‘ก,๐‘ข)โˆˆ0,2๐œ‹ร—(0,+โˆž),(3.2)(๐ป2)there exists a positive number ๐‘Ÿ>0 such that ๐ด2๎‚ปmax[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎€œ๐œ‰(๐‘ฅ)02๐œ‹๐‘(๐‘ )๐‘‘๐‘ +max[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎€œ๐œ‚(๐‘ฅ)02๐œ‹๎‚ผ๐‘ž(๐‘ )๐‘‘๐‘ +๐ด๐‘™๐›พ(๐‘Ÿ)<๐‘Ÿ,(3.3)(๐ป3)for the constant ๐‘Ÿ in (H2), there exists a function ฮฆ๐‘Ÿ>0 such that ๐‘“(๐‘ก,๐‘ข)>ฮฆ๐‘Ÿ[]],๎€œ(๐‘ก),(๐‘ก,๐‘ข)โˆˆ0,2๐œ‹ร—(0,๐‘Ÿ02๐œ‹ฮฆ๐‘Ÿ(๐‘ )๐‘‘๐‘ >0.(3.4) Then PBVP (1.1) has at least one positive periodic solution with 0<โ€–๐‘ขโ€–<๐‘Ÿ, where ๐‘’๐ด=2๐‘š๐œ‹+1๐‘š๎€ท๐‘’2๐‘š๐œ‹๎€ธ=๐‘’โˆ’1โˆš2๐œ‹๐‘€+1โˆš๐‘€๎‚€๐‘’โˆš2๐œ‹๐‘€๎‚โˆ’1.(3.5)

Proof. The existence of positive solutions is proved by using the Leray-Schauder alternative principle given in Lemma 2.2. We divide the rather long proof into six steps. Step 1. From (3.3), we may choose ๐‘›0โˆˆ{1,2,โ€ฆ} such that ๐ด2๎‚ปmax[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎€œ๐œ‰(๐‘ฅ)02๐œ‹๐‘(๐‘ )๐‘‘๐‘ +max[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎€œ๐œ‚(๐‘ฅ)02๐œ‹๎‚ผ1๐‘ž(๐‘ )๐‘‘๐‘ +๐ด๐‘™๐›พ(๐‘Ÿ)+๐‘›0<๐‘Ÿ.(3.6) Let ๐‘0={๐‘›0,๐‘›0+1,โ€ฆ}. For ๐‘›โˆˆ๐‘0. We consider the family of equations โˆ’๐‘ข๎…ž๎…ž(๐‘ก)+๐‘€๐‘ข(๐‘ก)=๐œ†๐‘“๐‘›๐‘€(๐‘ก,๐‘ข(๐‘ก))+๐‘›,๐‘กโˆˆ๐ฝ๎…ž,||ฮ”๐‘ข๐‘ก=๐‘ก๐‘˜=๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ,โˆ’ฮ”๐‘ข๎…ž||๐‘ก=๐‘ก๐‘˜=๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๐‘ข๎€ธ๎€ธ,๐‘˜=1,2,โ€ฆ๐‘™,(0)=๐‘ข(2๐œ‹),๐‘ข๎…ž(0)=๐‘ข๎…ž(2๐œ‹),(3.7) where ๐œ†โˆˆ[0,1] and ๐‘“๐‘›๎‚ต๎‚†1(๐‘ก,๐‘ข)=๐‘“๐‘ก,max๐‘ข,๐‘›๎‚‡๎‚ถ[,(๐‘ก,๐‘ข)โˆˆ๐ฝร—0,+โˆž).(3.8) For every ๐œ† and ๐‘›โˆˆ๐‘0, define an operator as follows: ๐‘‡๐œ†,๐‘›๎€œ๐‘ข(๐‘ก)=๐œ†02๐œ‹๐บ(๐‘ก,๐‘ )๐‘“๐‘›(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ +๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ,๐‘ขโˆˆ๐พ.(3.9) Then, we may verify that ๐‘‡๐œ†,๐‘›โˆถ๐พโŸถ๐พiscompletelycontinuous.(3.10) To find a positive solution of (3.7) is equivalent to solve the following fixed point problem in ๐‘ƒ๐ถ[๐ฝ,๐‘…]: ๐‘ข=๐‘‡๐œ†,๐‘›1๐‘ข+๐‘›.(3.11) Let ฮฉ={๐‘ฅโˆˆ๐พโˆถโ€–๐‘ฅโ€–<๐‘Ÿ},(3.12) then ฮฉ is a relatively open subset of the convex set ๐พ.Step 2. We claim that any fixed point ๐‘ข of (3.11) for any ๐œ†โˆˆ[0,1) must satisfies โ€–๐‘ขโ€–โ‰ ๐‘Ÿ.
Otherwise, we assume that ๐‘ข is a solution of (3.11) for some ๐œ†โˆˆ[0,1) such that โ€–๐‘ขโ€–=๐‘Ÿ. Note that ๐‘“๐‘›(๐‘ก,๐‘ข)โ‰ฅ0. ๐‘ข(๐‘ก)โ‰ฅ1/๐‘› for all ๐‘กโˆˆ๐ฝ and ๐‘Ÿโ‰ฅ๐‘ข(๐‘ก)โ‰ฅ(1/๐‘›)+๐œŽโ€–๐‘ขโˆ’1/๐‘›โ€–. By the choice of ๐‘›0, 1/๐‘›โ‰ค1/๐‘›0<๐‘Ÿ. Hence, for all ๐‘กโˆˆ๐ฝ, we get 1๐‘Ÿโ‰ฅ๐‘ข(๐‘ก)โ‰ฅ๐‘›โ€–โ€–โ€–1+๐œŽ๐‘ขโˆ’๐‘›โ€–โ€–โ€–โ‰ฅ1๐‘›|||1+๐œŽโ€–๐‘ขโ€–โˆ’๐‘›|||โ‰ฅ1๐‘›๎‚€1+๐œŽ๐‘Ÿโˆ’๐‘›๎‚>๐œŽ๐‘Ÿ.(3.13) From (3.2), we have ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธโ‰คmax1โ‰ค๐‘˜โ‰ค๐‘™๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ท๐‘ข๎€ท๐‘ก๎€ธ๎€ธโ‰ค๐›พ๐‘˜๎€ธ๎€ธโ‰ค๐›พ(๐‘Ÿ).(3.14) Consequently, for any fixed point ๐‘ข of (3.11), by (3.8), (3.13), and (3.14), we have ๎€œ๐‘ข(๐‘ก)=๐œ†02๐œ‹๐บ(๐‘ก,๐‘ )๐‘“๐‘›(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ +๐‘˜=1๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+1๎€ธ๎€ธ๐‘›โ‰ค๎€œ02๐œ‹๐บ(๐‘ก,๐‘ )๐‘“(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ +๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+1๎€ธ๎€ธ๐‘›โ‰ค๎€œ02๐œ‹+1๐บ(๐‘ก,๐‘ )๐‘“(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ ฮ“๎ƒฏ๎“๐‘ก๐‘˜โ‰ค๐‘ก๎€บ๐‘’๐‘š(๐‘กโˆ’๐‘ก๐‘˜)+๐‘’๐‘š(2๐œ‹โˆ’๐‘ก+๐‘ก๐‘˜)๎€ป๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๐‘ก๐‘˜>๐‘ก๎€บ๐‘’๐‘š(๐‘ก๐‘˜โˆ’๐‘ก)+๐‘’๐‘š(2๐œ‹โˆ’๐‘ก๐‘˜+๐‘ก)๎€ป๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๐‘ก๐‘˜โ‰ค๐‘ก๎€บโˆ’๐‘’๐‘š(๐‘กโˆ’๐‘ก๐‘˜)+๐‘’๐‘š(2๐œ‹โˆ’๐‘ก+๐‘ก๐‘˜)๎€ป๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๐‘ก๐‘˜>๐‘ก๎€บ๐‘’๐‘š(๐‘ก๐‘˜โˆ’๐‘ก)โˆ’๐‘’๐‘š(2๐œ‹โˆ’๐‘ก๐‘˜+๐‘ก)๎€ป๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎ƒฐ+1๎€ธ๎€ธ๐‘›=๎€œ02๐œ‹+1๐บ(๐‘ก,๐‘ )๐‘“(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ ฮ“๎ƒฏ๎“๐‘ก๐‘˜โ‰ค๐‘ก๐‘’๐‘š(๐‘กโˆ’๐‘ก๐‘˜)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธโˆ’๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜โ‰ค๐‘ก๐‘’๐‘š(2๐œ‹โˆ’๐‘ก+๐‘ก๐‘˜)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ+๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜>๐‘ก๐‘’๐‘š(๐‘ก๐‘˜โˆ’๐‘ก)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ+๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜>๐‘ก๐‘’๐‘š(2๐œ‹โˆ’๐‘ก๐‘˜+๐‘ก)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธโˆ’๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎ƒฐ+1๎€ธ๎€ธ๎€ป๐‘›.(3.15) It follows from โˆ’(1/๐‘š)๐ฝ๐‘˜(๐‘ข)<๐ผ๐‘˜(๐‘ข)<(1/๐‘š)๐ฝ๐‘˜(๐‘ข) that ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธโˆ’๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ>0,๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ+๐‘š๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜๎€ธ๎€ธ>0.(3.16) So, we get from (3.1), (3.2), and (3.3) that ๎€œ๐‘ข(๐‘ก)โ‰ค02๐œ‹2๎€ท๐‘’๐บ(๐‘ก,๐‘ )๐‘“(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ +2๐‘š๐œ‹๎€ธ+1ฮ“๐‘™๎“๐‘˜=1๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+1๎€ธ๎€ธ๐‘›โ‰ค๎€œ02๐œ‹[]1๐บ(๐‘ก,๐‘ )๐‘(๐‘ )๐œ‰(๐‘ข(๐‘ ))+๐‘ž(๐‘ )๐œ‚(๐‘ข(๐‘ ))๐‘‘๐‘ +๐ด๐‘™๐›พ(๐‘Ÿ)+๐‘›0โ‰ค๐ด2๎‚ธ๎€œ02๐œ‹๐‘(๐‘ )๐‘‘๐‘ max[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎€œ๐œ‰(๐‘ฅ)+02๐œ‹๐‘ž(๐‘ )๐‘‘๐‘ max[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎‚น1๐œ‰(๐‘ฅ)+๐ด๐‘™๐›พ(๐‘Ÿ)+๐‘›0.(3.17) Therefore, ๐ด๐‘Ÿ=โ€–๐‘ขโ€–โ‰ค2๎‚ธ๎€œ02๐œ‹๐‘(๐‘ )๐‘‘๐‘ max[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎€œ๐œ‰(๐‘ฅ)+02๐œ‹๐‘ž(๐‘ )๐‘‘๐‘ max[]๐‘ฅโˆˆ๐œŽ๐‘Ÿ,๐‘Ÿ๎‚น1๐œ‰(๐‘ฅ)+๐ด๐‘™๐›พ(๐‘Ÿ)+๐‘›0<๐‘Ÿ.(3.18) This is a contraction, and so the claim is proved.
Step 3. From the above claim and the Leray-Schauder alternative principle, we know that operator (3.9) (with ๐œ†=1) has a fixed point denoted by ๐‘ข๐‘› in ฮฉ. So, (3.7) (with ๐œ†=1) has a positive solution ๐‘ข๐‘› with โ€–โ€–๐‘ข๐‘›โ€–โ€–<๐‘Ÿ,๐‘ข๐‘›1(๐‘ก)โ‰ฅ๐‘›,๐‘กโˆˆ๐ฝ.(3.19)Step 4. We show that {๐‘ข๐‘›} have a uniform positive lower bound; that is, there exists a constant ๐›ฟ>0, independent of ๐‘›โˆˆ๐‘0, such that min๐‘ก๎€ฝ๐‘ข๐‘›๎€พ(๐‘ก)โ‰ฅ๐›ฟ.(3.20) In fact, from (3.4), (3.8), (3.16), and (3.19), we get ๐‘ข๐‘›๎€œ(๐‘ก)=02๐œ‹๐บ(๐‘ก,๐‘ )๐‘“๐‘›๎€ท๐‘ ,๐‘ข๐‘›๎€ธ(๐‘ )๐‘‘๐‘ +๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+1๎€ธ๎€ธ๐‘›=๎€œ02๐œ‹๎€ท๐บ(๐‘ก,๐‘ )๐‘“๐‘ ,๐‘ข๐‘›๎€ธ(๐‘ )๐‘‘๐‘ +๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+1๎€ธ๎€ธ๐‘›โ‰ฅ๎€œ02๐œ‹๐บ(๐‘ก,๐‘ )ฮฆ๐‘Ÿ+1(๐‘ )๐‘‘๐‘ ฮ“๎ƒฏ๎“๐‘ก๐‘˜โ‰ค๐‘ก๐‘’๐‘š(๐‘กโˆ’๐‘ก๐‘˜)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธโˆ’๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜โ‰ค๐‘ก๐‘’๐‘š(2๐œ‹โˆ’๐‘ก+๐‘ก๐‘˜)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธ+๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜>๐‘ก๐‘’๐‘š(๐‘ก๐‘˜โˆ’๐‘ก)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธ+๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜>๐‘ก๐‘’๐‘š(2๐œ‹โˆ’๐‘ก๐‘˜+๐‘ก)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธโˆ’๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎ƒฐ+1๎€ธ๎€ธ๎€ป๐‘›โ‰ฅ๎€œ02๐œ‹๐บ(๐‘ก,๐‘ )ฮฆ๐‘Ÿโ‰ฅ(๐‘ )๐‘‘๐‘ 2๐‘’๐‘š๐œ‹ฮ“๎€œ02๐œ‹ฮฆ๐‘Ÿ(๐‘ )๐‘‘๐‘ โˆถ=๐›ฟ>0.(3.21)Step 5. We prove that โ€–โ€–๐‘ข๎…ž๐‘›โ€–โ€–<๐ป,๐‘›โ‰ฅ๐‘›0(3.22) for some constant ๐ป>0. Equations (3.19) and (3.20) tell us that ๐›ฟโ‰ค๐‘ข๐‘›(๐‘ก)โ‰ค๐‘Ÿ, so we may let ๐‘€1=max[]๐‘กโˆˆ๐ฝ,๐‘ขโˆˆ๐›ฟ,๐‘Ÿ๐‘“(๐‘ก,๐‘ข),๐‘€2=max๐‘ก,๐‘ โˆˆ๐ฝ||๐บ๎…ž๐‘ก||(๐‘ก,๐‘ ),๐‘€3=max[]๐‘™๐‘ขโˆˆ๐›ฟ,๐‘Ÿ๎“๐‘˜=1๐ฝ๐‘˜(๐‘ข).(3.23) Then, โ€–โ€–๐‘ข๎…ž๐‘›โ€–โ€–=sup๐‘กโˆˆ๐ฝ||๐‘ข๎…ž๐‘›||(๐‘ก)=sup๐‘กโˆˆ๐ฝ|||||๎€œ02๐œ‹๐บ๎…ž๐‘ก๎€ท(๐‘ก,๐‘ )๐‘“๐‘ ,๐‘ข๐‘›๎€ธ(๐‘ )๐‘‘๐‘ +๐‘™๎“๐‘˜=1๐บ๎…ž๐‘ก๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๎‚ต๐œ•๐‘ก๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๎‚ถ๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜||||๎€ธ๎€ธ=sup๐‘กโˆˆ๐ฝ||||๎€œ02๐œ‹๐บ๎…ž๐‘ก๎€ท(๐‘ก,๐‘ )๐‘“๐‘ ,๐‘ข๐‘›๎€ธ+๐‘š(๐‘ )๐‘‘๐‘ ฮ“๎ƒฏ๎“๐‘ก๐‘˜โ‰ค๐‘ก๐‘’๐‘š(๐‘กโˆ’๐‘ก๐‘˜)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธโˆ’๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜โˆ’๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜โ‰ค๐‘ก๐‘’๐‘š(2๐œ‹โˆ’๐‘ก+๐‘ก๐‘˜)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธ+๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜โˆ’๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜>๐‘ก๐‘’๐‘š(๐‘ก๐‘˜โˆ’๐‘ก)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธ+๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜+๎“๎€ธ๎€ธ๎€ป๐‘ก๐‘˜>๐‘ก๐‘’๐‘š(2๐œ‹โˆ’๐‘ก๐‘˜+๐‘ก)๎€บ๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธโˆ’๐‘š๐ผ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎ƒฐ|||||๎€ธ๎€ธ๎€ปโ‰คsup๐‘กโˆˆ๐ฝ๎€œ02๐œ‹||๐บ๎…ž๐‘ก||๐‘“๎€ท(๐‘ก,๐‘ )๐‘ ,๐‘ข๐‘›๎€ธ๎€ท๐‘’(๐‘ )๐‘‘๐‘ +2๐‘š2๐‘š๐œ‹๎€ธ+1ฮ“๐‘™๎“๐‘˜=1๐ฝ๐‘˜๎€ท๐‘ข๐‘›๎€ท๐‘ก๐‘˜๎€ธ๎€ธโ‰ค2๐œ‹๐‘€1๐‘€2+๎€ท๐‘’2๐‘š2๐‘š๐œ‹๎€ธ+1ฮ“๐‘€3โˆถ=๐ป.(3.24)Step 6. Now, we pass the solution ๐‘ข๐‘› of the truncation equation (3.7) (with ๐œ†=1) to that of the original equation (1.1). The fact that โ€–๐‘ข๐‘›โ€–<๐‘Ÿ and (3.22) show that {๐‘ข๐‘›}๐‘›โˆˆ๐‘0 is a bounded and equi-continuous family on [0,2๐œ‹]. Then, the Arzela-Ascoli Theorem guarantees that {๐‘ข๐‘›}๐‘›โˆˆ๐‘0 has a subsequence {๐‘ข๐‘›๐‘—}๐‘—โˆˆ๐‘, converging uniformly on [0,2๐œ‹]. From the fact โ€–๐‘ข๐‘›โ€–<๐‘Ÿ and (3.20), ๐‘ข satisfies ๐›ฟโ‰ค๐‘ข(๐‘ก)โ‰ค๐‘Ÿ for all ๐‘กโˆˆ๐ฝ. Moreover, ๐‘ข๐‘›๐‘— also satisfies the following integral equation: ๐‘ข๐‘›๐‘—๎€œ(๐‘ก)=02๐œ‹๎‚€๐บ(๐‘ก,๐‘ )๐‘“๐‘ ,๐‘ข๐‘›๐‘—๎‚+(๐‘ )๐‘‘๐‘ ๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎‚€๐‘ข๐‘›๐‘—๎€ท๐‘ก๐‘˜๎€ธ๎‚+๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎‚€๐‘ข๐‘›๐‘—๎€ท๐‘ก๐‘˜๎€ธ๎‚+1๐‘›๐‘—.(3.25) Let ๐‘—โ†’+โˆž, and we get ๎€œ๐‘ข(๐‘ก)=02๐œ‹+๐บ(๐‘ก,๐‘ )๐‘“(๐‘ ,๐‘ข(๐‘ ))๐‘‘๐‘ ๐‘™๎“๐‘˜=1๐บ๎€ท๐‘ก,๐‘ก๐‘˜๎€ธ๐ฝ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜+๎€ธ๎€ธ๐‘™๎“๐‘˜=1๐œ•๐บ(๐‘ก,๐‘ )|||๐œ•๐‘ ๐‘ =๐‘ก๐‘˜๐ผ๐‘˜๎€ท๐‘ข๎€ท๐‘ก๐‘˜,๎€ธ๎€ธ(3.26) where the uniform continuity of ๐‘“(๐‘ก,๐‘ข) on ๐ฝร—[๐›ฟ,๐‘Ÿ] is used. Therefore, ๐‘ข is a positive solution of PBVP (1.1). This ends the proof.

4. An Example

Consider the following impulsive PBVP:โˆ’๐‘ข๎…ž๎…ž(๐‘ก)+๐‘€๐‘ข(๐‘ก)=๐‘ก2๎‚ต||||1+sin๐‘ข๐‘ข3/2๎‚ถ+๐‘ก(1+|cos๐‘ข|),๐‘กโˆˆ๐ฝ๎…ž,||ฮ”๐‘ข๐‘ก=๐‘ก๐‘˜=๎€ฝ๐‘min1,๐‘2,โ€ฆ,๐‘๐‘™๎€พ2โˆš๐‘€๐‘ข๎€ท๐‘ก๐‘˜๎€ธ,โˆ’ฮ”๐‘ข๎…ž||๐‘ก=๐‘ก๐‘˜=๐‘๐‘˜๐‘ข๎€ท๐‘ก๐‘˜๎€ธ,๐‘˜=1,2,โ€ฆ,๐‘™,๐‘ข(0)=๐‘ข(2๐œ‹),๐‘ข๎…ž(0)=๐‘ข๎…ž(2๐œ‹),(4.1) where ๐‘๐‘˜>0 are constants. Then, PBVP (4.1) has at least one positive solution ๐‘ข with 0<โ€–๐‘ขโ€–<1.

To see this, we will apply Theorem 3.1.

Let ๐‘“(๐‘ก,๐‘ข)=๐‘ก2๎‚ต||||1+sin๐‘ข๐‘ข3/2๎‚ถ+๐‘ก(1+|cos๐‘ข|),(4.2) then ๐‘“(๐‘ก,๐‘ข) has a repulsive singularity at ๐‘ข=0lim๐‘ขโ†’0+๐‘“(๐‘ก,๐‘ข)=+โˆž,uniformalyin๐‘ก.(4.3) Denote ๐‘(๐‘ก)=๐‘ก2||||,๐‘ž(๐‘ก)=๐‘ก,๐œ‰(๐‘ข)=1+sin๐‘ข๐‘ข3/2๎€ฝ๐‘,๐œ‚(๐‘ข)=1+|cos๐‘ข|,๐›พ(๐‘ข)=max1,๐‘2,โ€ฆ,๐‘๐‘™๎€พ๐‘ข,๐‘Ÿ=1,ฮฆ๐‘Ÿ(๐‘ก)=๐‘ก+๐‘ก2.(4.4) Then, it is easy to say that (3.1), (3.2), and (3.3) hold. From (3.5), we knowlim๐‘€โ†’+โˆž๐ด=lim๐‘€โ†’+โˆž๐‘’โˆš2๐œ‹๐‘€+1โˆš๐‘€๎‚€๐‘’โˆš2๐œ‹๐‘€๎‚โˆ’1=0.(4.5) So, we may choose ๐‘€ large enough to guarantee that (3.3) holds. Then, the result follows from Theorem 3.1.

Remark 4.1. Functions ๐œ‰, ๐œ‚ in example (4.1) do not have the monotonicity required as in [19]. So, the results of [19] cannot be applied to PBVP (4.1).

Acknowledgments

The authors are grateful to the anonymous referees for their helpful suggestions and comments. Zhaocai Hao acknowledges support from NSFC (10771117), Ph.D. Programs Foundation of Ministry of Education of China (20093705120002), NSF of Shandong Province of China (Y2008A24), China Postdoctoral Science Foundation (20090451290), ShanDong Province Postdoctoral Foundation (200801001), and Foundation of Qufu Normal University (BSQD07026).