• No results found

Periodic Boundary Value Problems for Second Order Functional Differential Equations

N/A
N/A
Protected

Academic year: 2020

Share "Periodic Boundary Value Problems for Second Order Functional Differential Equations"

Copied!
11
0
0

Loading.... (view fulltext now)

Full text

(1)

Volume 2010, Article ID 598405,11pages doi:10.1155/2010/598405

Research Article

Periodic Boundary Value Problems for

Second-Order Functional Differential Equations

Xiangling Fu and Weibing Wang

Department of Mathematics, Hunan University of Science and Technology, Xiangtan, Hunan 411201, China

Correspondence should be addressed to Xiangling Fu,[email protected] Received 11 June 2009; Accepted 28 January 2010

Academic Editor: Marta Garc´ıa-Huidobro

Copyrightq2010 X. Fu and W. Wang. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This note considers a periodic boundary value problem for a second-order functional differential equation. We extend the concept of lower and upper solutions and obtain the existence of extreme solutions by using upper and lower solution method.

1. Introduction

Upper and lower solution method plays an important role in studying boundary value problems for nonlinear differential equations; see1 and the references therein. Recently, many authors are devoted to extend its applications to boundary value problems of functional differential equations2–5. Supposeαis one upper solution or lower solution of periodic boundary value problems for second-order differential equation; the condition

α0 αTis required. A neutral problem is that whether we can define upper and lower solution without assuming α0 αT. The aim of the present paper is to discuss the following second order functional differential equation

ut ft, ut, uθt, tJ,

u0 uT, u0 uT, 1.1

whereJ 0, T, f∈CJ×R2, R,0θtt, tJ.

(2)

Through this paper, we assume thatM >0, N≥0,and

T2MN≤2, 1.2

Λ1

uC2J:ut≥0,1u0 uT, u0≥uT,

Λ2

uC2J:ut≥0,1uT u0, u0≤uT.

1.3

Definition 1.1. FunctionsαC2JandβC2Jare called lower solution and upper solution

of the boundary value problem1.1, respectively if

αt≤ft, αt, αθtt, tJ,

α0≥αT, 1.4

βt≥ft, βt, βθthβt, tJ,

β0≤βT, 1.5

where

t

⎧ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎩

0, α0 αT,

ct Mct Ncθtα0−αT, α0> αT,

bt Mbt NbθtαTα0, α0< αT,

1.6

t

⎧ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎩

0, β0 βT,

ct Mct Ncθtβ0−βT, β0< βT,

bt Mbt NbθtβTβ0, β0> βT

1.7

andc∈Λ1, b∈Λ2.

Remark 1.2. Clearly,Λ1/∅andΛ2∅. For example,

t

T ∈Λ1, sin πt

2T ∈Λ1, 1− t

T ∈Λ2, sin π

2

1− t

T

(3)

2. Comparison Results

We now present the main results of this section.

Lemma 2.1. Assume thatuC2Jsatisfies

ut Mut Nuθt≤0, tJ,

u0 uT, u0≥uT, 2.1

thenut≤0for alltJ.

Proof. Suppose, to the contrary, thatut >0 for sometJ. We consider the following two cases.

Case 1. ut≥0, ut/0 onJ. It is easy to obtain thatut≥0 onJ. Thusutconstant K >0 fromu0≥uT. Consequently, we obtain that

MNKut Mut Nuθt≤0, 2.2

which contradictsK >0.

Case 2. There existt1, t2∈Jsuch thatut1>0 andut2<0. Hence, two cases are possible.

Subcase 1. u0 uT<0. There exists at3∈0, Tsuch that

ut3 max

tJ ut>0, u

t

3 0. 2.3

Letut∗ mint∈0,t3ut<0.Then

ut≥MNut∗, t∈0, t3. 2.4

Integrating the above inequality fromstst3tot3, we obtain

us≥t3−sMNut∗, t∗≤st3, 2.5

and then integrate fromt∗tot3to obtain

ut< ut3−ut∗≤

t3

t

s−t3MNut∗ds

≤ −MN

2 ut∗t3−t

2≤ −MN

2 ut∗T

2

2.6

(4)

Subcase 2. u0 uT≥0. There exists at3∈Jsuch that

ut3 max

tJ ut>0. 2.7

Ift3∈0, T, thenut3 0.Ift30 ort3T, thenu0≤0≤uT. Sou0 uT 0.

Letut∗ mint∈0,Tut<0.Then

ut≥MNut∗, tJ. 2.8

Whent< t3, same as Subcase 1, we obtain that 1< T2MN/2.

whent> t3, integrating the inequality2.8fromt3tost3≤st∗, we obtain

us≥s−t3MNut∗, 2.9

and then integrate fromt3tot∗to obtain

ut> ut∗−ut3≥

t

t3

s−t3MNut∗ds≥ M2Nut∗t∗−t32≥ MN

2 ut∗T

2

2.10

that implies 1<T2/2MN.This is a contradiction. The proof is complete.

Lemma 2.2. Assume thatuC2Jsatisfies

ut Mut Nuθt ct Mct Ncθtu0−uT≤0, tJ, u0> uT, u0≥uT.

2.11

Thenut≤0for alltJ.

Proof. Put

yt ut ctu0−uT, tJ, 2.12

thenyEandutytfor alltJand

(5)

We have

yt Myt Nyθt ut Mut Nuθt

ct Mct Ncθtu0−uT≤0,

y0 1c0u0−c0uT yT,

y0 u0 c0u0−uTyT.

2.14

Hence by Lemma 2.1,yt ≤ 0 for all tJ, which implies that ut ≤ 0 for tJ. This completes the proof.

Lemma 2.3. Assume thatuC2Jsatisfies

ut Mut Nuθt bt Mbt NbθtuT−u0≤0, tJ,

u0< uT, u0≥uT. 2.15

Thenut≤0for alltJ.

The proof ofLemma 2.3is similar to that ofLemma 2.2, here we omit it.

Lemma 2.4. Assume thatuC2Jsatisfies

ut Mut Nuθt≤0, tJ,

u0≤0, uT≤0. 2.16

Thenut≤0for alltJ.

Proof. Suppose thatut > 0 for sometJ. Then from boundary conditions, we have that there exists at∗∈0,1such that

utmax

tJ ut>0, u

t0.

2.17

Suppose thatut ≥0 fortJ. It is easy to see thatu0 0 andut ≥0 fortJ. Fromu0 0 andut≥0 fortJ, we obtain thatu0≥0. Therefore,ut≥u0≥0. It follows thatuT maxtJut>0, a contradiction.

Suppose that there existt1, t2 ∈Jsuch thatut1 >0 andut2<0. Lett∗ ∈0, t∗be

such thatut∗ mint∈0,tut≤0. From the first inequality of2.16, we have

ut≥MNut∗, t∈0, t∗. 2.18

Integrating the above inequality fromst∗≤st∗tot∗, we obtain

(6)

and then integrate fromt∗tot∗to obtain

ut< ut∗−ut∗≤ t

t

s−t∗MNut∗ds

≤ −MN

2 ut∗t

t

∗2≤ −M2Nut∗T2.

2.20

Hence,ut∗2−MNT2>0,a contradiction. The proof is complete.

3. Linear Problem

In this section, we consider the boundary value problem

ut Mut Nuθt σt, tJ,

u0 uT, u0 uT,

3.1

whereσCJ, R.

Theorem 3.1. Assume that there existα, βwhich are lower and upper solutions of 3.1andαβ

onJ. Then there exists one unique solutionuto problem3.1andαuβonJ.

Proof. We first show that the solution of3.1is unique. Letu1andu2be solutions of3.1and

setvu1−u2. Thus

vt Mvt Nvθt 0, tJ,

v0 vT, v0 vT.

3.2

ByLemma 2.1, we have thatv≤0 fortJ, that is,u1≤u2onJ. Similarly, one can obtain that

u2≤u1onJ. Henceu1 u2.

Next, we prove that ifuis a solution of3.1, thenαuβ. Letu. Ifα0 αT, thent 0 onJ. So we have

mt Mmt Nmθt≤0, tJ,

m0 mT, m0≥mT.

3.3

ByLemma 2.1, we have thatu≤0 onJ.

Ifα0> αT, thent −ct Mct Ncθtα0−αT. Thus

mt Mmt Nmθt αt Mαt Nαθt

ut−MutNuθt

σtt−σt t

−−ct Mct Ncθtm0−mT.

(7)

It is easy to see that m0 ≥ mT.By Lemma 2.2, we have that m αu ≤ 0 on J. Analogously,uβonJ.

Ifα0< αT, thent −bt Mbt NbθtαTα0. Thus

mt Mmt Nmθt αt Mαt Nαθt

ut−MutNuθt

σtt−σt t

−−bt Mbt NbθtmT−m0.

3.5

It is easy to see that m0 ≥ mT.By Lemma 2.2, we have that m αu ≤ 0 on J. Analogously,uβonJ.

Finally, we show that3.1has a solution by several steps.

Step 1. Consider the equation

ut Mut Nuθt σt, tJ,

u0 uT λ, 3.6

where M, N and σ are defined in3.1. For any λR, we show that3.6 has a unique solutionut, λ.

It is easy to check that3.6is equivalent to the integral equation

ut λ

T

0

Gt, sσsMusNuθsds, 3.7

where

Gt, s

⎧ ⎪ ⎨ ⎪ ⎩

t

TT−s, 0≤tsT, s

TT−t, 0≤stT.

3.8

Define a mappingΦ:CJ, RCJ, Rby

Φut λ

T

0

Gt, sσsMusNuθsds. 3.9

ObviouslyCJ, Ris a Banach space with normu suptJ|ut|. For anyx, yCJ, R, we have

Φxt−Φyt

T

0

(8)

Noting that maxtJ

T

0Gt, sds T2/8, condition1.2implies thatΦ :CJ, RCJ, Ris

a contraction mapping. There exists uniqueuCJ, Rsuch thatΦu u. Thus3.6has a unique solutionut, λ.Moreover,ut, λC2J, R.

Step 2. We show that for anytJ,ut, λandutt, λare continuous inλ, whereut, λis a

unique solution of the problem3.6. Letut, λi, i1,2,be the solution of

ut Mut Nuθt σt, tJ, u0 uT λi, i1,2.

3.11

Then

ut, λi λi

T

0

Gt, sσsMus, λiNuθs, λids, i1,2. 3.12

From3.12, we have that

ut, λ1−ut, λ2 ≤ |λ1−λ2| MNut, λ1−ut, λ2max

tJ

T

0

Gt, sds

≤ |λ1−λ2|T 2

8 MNut, λ1−ut, λ2.

3.13

Hence

ut, λ1−ut, λ2 ≤

8

8−T2MN|λ1−λ2|. 3.14

Sinceut, λC2J, R,u

tt, λexists for anytJandλR. From3.12and3.14, we have

utt, λ1−utt, λ2 ≤ MN

T ut, λ1−ut, λ2maxtJ

t

0

s ds

T

t

T−sds

≤ MNT

2 ut, λ1−ut, λ2

≤ 4MNT

8−T2MN|λ1−λ2|.

(9)

Step 3. We show that there exists oneλsuch thatut0, λ utT, λ, whereut, λis the unique

solution of the problem3.6. Put

αt

⎧ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎩

αt, α0 αT,

αt ctα0−αT, α0> αT,

αt btαTα0, α0< αT,

3.16

βt

⎧ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎩

βt, β0 βT,

βtctβTβ0, β0< βT,

βtbtβ0−βT, β0> βT;

3.17

thenαtαt, βtβtfor anytJand

αt Mαt Nαθtδt, tJ,

α0 αT, α0≥αT,

3.18

βt Mβt Nβθtδt, tJ,

β0 βT, β0≤βT.

3.19

Using3.18and3.19, one easily obtains thatαtβtfor anytJ.

Put λ ∈ α0, β0, then α0 αTu0, λ uT, λβ0 βT.Using

Lemma 2.3, we easily obtain thatαtut, λβtonJ. Hence

ut0, α0≥α0, utT, αT≤αT, 3.20

ut

0, β0≤β0, ut

T, βTβT. 3.21

Define a function

Pλ ut0, λutT, λ, 3.22

whereut, λis the unique solution of the problem3.6. SincePis continuous and

Pα0Pβ0≤0, 3.23

there exists oneλ0 ∈α0, β0such that0 0,that is,ut0, λ0 utT, λ0. Obviously,

(10)

4. Main Result

Theorem 4.1. Let the following conditions hold.

H1The functionsα,βare lower and upper solutions of1.1, respectively, andαβonJ.

H2The constantsM, Nin definition of upper and lower solutions satisfy

ft, x, yft, x, y≥ −MxxNyy 4.1

forαtxxβt, αθtyyβθt, tJ.

Then, there exist monotone sequences {αn}, {βn} with α0 α, β0 β such that

limn→ ∞αnt ρt,limn→ ∞βnt rtuniformly onJ, andρ, rare the minimal and the maximal

solutions of 1.1, respectively, such that

α0≤α1≤α2≤ · · · ≤αnρxrβn≤ · · · ≤β2≤β1≤β0 4.2

onJ, wherexis any solution of 1.1such thatαtxtβtonJ.

Proof. Letα, β {uC2J: αt utβt, t J}. For anyγ α, β, we consider the

equation

ut Mut Nuθt ft, γt, γθtMγt Nγθt, tJ,

u0 uT, u0 uT. 4.3

Theorem 3.1implies that the problem4.3has a unique solutionuC2J. We define an

operatorAbyuAγ, thenAis an operator fromα, βtoα, β. We shall show that

aαAα, β;

bAis nondecreasing inα, β.

From∈α, βand∈α, β, we have thataholds. To proveb, we show that

1≤2ifαμ1≤μ2≤β.

Letρ11, ρ22,and1ρ2; then byH2, we have

pt Mpt Npθtft, μ1t, μ1θt

1t 1θt

ft, μ2t, μ2θt

2t−2θt

≤0,

4.4

andp0 pT, p0 pT.ByLemma 2.1,pt≤0, which implies1≤2.

Define the sequence{αn}, {βn}withα0 α, β0 βsuch thatαn1 Aαn, βn1 Aβn

forn0,1,2, . . . .Fromaandb, we have

(11)

on tJ. Therefore, there exist ρ, r such that limn→ ∞αnt ρt, limn→ ∞βnt rt

uniformly onJ. Clearly,ρ, rare solutions of1.1.

Finally, we prove that ifx∈α0, β0is one solution of1.1, thenρtxtrton

J. Sinceα0t≤xtβ0tandAxx, by propertyb, we obtain thatα1t≤xtβ1t

fortJ. Using propertybrepeatedly, we have

αnt≤xtβnt 4.6

for alln. Passing to the limit asn → ∞, we obtainρtxtrt, tJ.This completes the proof.

Acknowledgment

This work is supported by Scientific Research Fund of Hunan Provincial Education Department09B033and the NNSF of China10871062.

References

1 G. S. Ladde, V. Lakshmikantham, and A. S. Vatsala, Monotone Iterative Techniques for Nonlinear Differential Equations, Pitman, London, UK, 1985.

2 J. J. Nieto and R. Rodriguez-Lopez, “Existence and approximation of solutions for nonlinear function differential equations with periodic boundary value conditions,”Journal of Computational and Applied Mathematics, vol. 40, pp. 433–442, 2000.

3 J. J. Nieto and R. Rodriguez-Lopez, “Remark on periodic boundary value problems for function differential equations,”Journal of Computational and Applied Mathematics, vol. 158, pp. 339–353, 2003.

4 Z. He and X. He, “Periodic boundary value problems for first order impulsive integro-differential equations of mixed type,”Journal of Mathematical Analysis and Applications, vol. 296, no. 1, pp. 8–20, 2004.

References

Related documents

Furthermore, it is expected that the idea of imposing rank minimization constraint on tensor latent space can be extended to various tensor decomposition mod- els in order to

Interpretation and analysis of the medical images constitute the important and major part of the detection field of a machine. Computer- aided detection systems are used for

Table 2 illustrates the average score of participants in different aspects of the ability to take medical history, showing that the aspect with the highest point was

Methods: It was a descriptive (combined retrospective and prospective) study of all patients with abdominal trauma admitted and managed at Ekiti State University Teaching

As definition of the services REST, for both the sets of points and the possible triangulations, was based on the Entity Class associated with the persistence provided

We first express the multivariable gimel-function by this multiple integrals contour with the help of (1.1), express the general class of mutivariable polynomials occurring on

 bully agents, when they send too many messages.  mistreated agents, when they received too many messages.  mistreated-bully agents, when they send and receive too

We believe that the national (economic) security of the country and its regions should be based on a geofinancial balance - an analytical and synthetic tool that reflects the