• 検索結果がありません。

and Its Application to Periodic Solutions of Integrodifferential Impulsive Periodic

N/A
N/A
Protected

Academic year: 2022

シェア "and Its Application to Periodic Solutions of Integrodifferential Impulsive Periodic"

Copied!
22
0
0

読み込み中.... (全文を見る)

全文

(1)

Volume 2008, Article ID 430521,22pages doi:10.1155/2008/430521

Research Article

The Generalized Gronwall Inequality

and Its Application to Periodic Solutions of Integrodifferential Impulsive Periodic

System on Banach Space

JinRong Wang,1X. Xiang,1, 2W. Wei,2and Qian Chen2

1College of Computer Science and Technology, Guizhou University, Guiyang, Guizhou 550025, China

2College of Science, Guizhou University, Guiyang, Guizhou 550025, China

Correspondence should be addressed to JinRong Wang,wjr9668@126.com Received 27 June 2008; Accepted 29 September 2008

Recommended by Ondˇrej Doˇsl ´y

This paper deals with a class of integrodifferential impulsive periodic systems on Banach space.

Using impulsive periodic evolution operator given by us, theT0-periodic PC-mild solution is introduced and suitable Poincar´e operator is constructed. Showing the compactness of Poincar´e operator and using a new generalized Gronwall’s inequality with impulse, mixed type integral operators andB-norm given by us, we utilize Leray-Schauder fixed point theorem to prove the existence ofT0-periodic PC-mild solutions. Our method is much different from methods of other papers. At last, an example is given for demonstration.

Copyrightq2008 JinRong Wang et al. 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.

1. Introduction

It is well known that impulsive periodic motion is a very important and special phenomenon not only in natural science, but also in social science such as climate, food supplement, insecticide population, and sustainable development. Periodic system with applications on finite-dimensional spaces has been extensively studied. Particularly, impulsive periodic systems on finite-dimensional spaces are considered and some important resultssuch as the existence and stability of periodic solution, the relationship between bounded solution and periodic solution, and robustness by perturbationare obtainedsee1–4.

Since the end of last century, many researchers pay great attention to impulsive systems on infinite-dimensional spaces. Particulary, Ahmed et al. investigated optimal control problems of system governed by impulsive system see 5–8. Many authors including us also gave a series of results for semilinear integrodifferential, strongly nonlinearimpulsive systems and optimal control problemssee9–20.

(2)

Although, there are some papers on periodic solution for periodic system on infinite- dimensional spacessee12,21–23and some results discussing integrodifferential system on finite Banach space and infinite Banach space see 11, 13. To our knowledge, inte- grodifferential impulsive periodic systems on infinite-dimensional spaceswith unbounded operator have not been extensively investigated. Recently, we discuss the impulsive periodic system and integrodifferential impulsive system on infinite-dimensional spaces.

Linear impulsive evolution operator is constructed and T0-periodic P C-mild solution is introduced. The existence of periodic solutions, alternative theorem criteria of Massera type, asymptotical stability, and robustness by perturbation is established see 24–26.

For semilinear impulsive periodic system, a suitable Poincar´e operator is constructed which verifies its compactness and continuity. By virtue of a generalized Gronwall inequality with mixed integral operator and impulse given by us, the estimate of theP C-mild solutions is derived. Some fixed point theorems such as Banach fixed point theorem and Horn fixed point theorem are applied to obtain the existence of periodicP C-mild solutions, respectivelysee 27,28. For integrodifferential impulsive system, the existence ofP C-mild solutions and optimal controls is presentedsee15.

Herein, we go on studying the following integrodifferential impulsive periodic system

xt ˙ Axt f

t, x, t

0

gt, s, xds

, t /τk, Δxt Bkxt ck, k.

1.1

on infinite-dimensional Banach spaceX, where 0τ0 < τ1< τ2<· · ·< τk· · ·; limk→ ∞τk ∞, τ τk T0;Δxτk kk,k ∈ Z0;T0 is a fixed positive number; andδ ∈ N denoted the number of impulsive points between 0 andT0. The operatorAis the infinitesimal generator of aC0-semigroup{Tt, t ≥ 0}onX;f is aT0-periodic, with respect tot ∈ 0

∞, Carath´edory function; g is a continuous function from0,∞×0,∞×X toX and is T0-periodic int ands; and B Bk, c ck. This paper is mainly concerned with the existence of periodic solutions for integrodifferential impulsive periodic system on infinite- dimensional Banach spaceX.

In this paper, we use Leray-Schauder fixed point theorem to obtain the existence of periodic solutions for integrodifferential impulsive periodic system 1.1. First, by virtue of impulsive evolution operator corresponding to linear homogeneous impulsive system, we construct a new Poincar´e operator P for integrodifferential impulsive periodic system 1.1, then we overcome some difficulties to show the compactness of Poincar´e operatorP which is very important. By a new generalized Gronwall inequality with impulse, mixed- type integral operators, andB-norm given by us, the estimate of fixed point set{xλP x, λ∈ 0,1}is established. Therefore, the existence ofT0-periodicP C-mild solutions for impulsive integrodifferential periodic system is shown.

In order to obtain the existence of periodic solutions, many authors use Horn fixed point theorem or Banach fixed point theorem. However, the conditions for Horn fixed point theorem are not easy to be verified sometimes and the conditions for Banach fixed point theorem are too strong. Our method is much different from others’, and we give a new way to show the existence of periodic solutions. In addition, the new generalized Gronwall inequality with impulse, mixed-type integral operator, andB-norm given by us, which can be used in other problems, have played an essential role in the study of nonlinear problems on infinite-dimensional spaces.

(3)

This paper is organized as follows. In Section 2, some results of linear impulsive periodic system and properties of impulsive periodic evolution operator corresponding to homogeneous linear impulsive periodic system are recalled. InSection 3, the new generalized Gronwall inequality with impulse, mixed-type integral operator, andB-norm are established.

In Section 4, the T0-periodic P C-mild solution for integrodifferential impulsive periodic system 1.1 is introduced. We construct the suitable Poincar´e operator P and give the relation betweenT0-periodicP C-mild solution and the fixed point ofP. After showing the compactness of the Poincar´e operatorPand obtaining the boundedness of the fixed point set {x λP x, λ ∈ 0,1}by virtue of the generalized Gronwall inequality, we can use Leray- Schauder fixed point theorem to establish the existence ofT0-periodicP C-mild solutions for integrodifferential impulsive periodic system. At last, an example is given to demonstrate the applicability of our result.

2. Linear impulsive periodic system

In order to study the integrodifferential impulse periodic system, we first recall some results about linear impulse periodic system here. LetXbe a Banach space. £Xdenotes the space of linear operators inX; £bXdenotes the space of bounded linear operators inX. £bXis the Banach space with the usual supremum norm. DefineD1, . . . , τδ} ⊂0, T0, whereδ∈N denotes the number of impulsive points between0, T0. We introduceP C0, T0;X≡ {x: 0, T0X|xto be continuous att∈0, T0\D; xis continuous from left and has right-hand limits attD}; and P C10, T0;X≡ {x∈P C0, T0;X|x˙ ∈P C0, T0;X}.Set

xP Cmax

sup

t∈0,T0

xt0, sup

t∈0,T0

xt−0

, xP C1xP Cx˙ P C. 2.1

It can be seen that endowed with the norm·P C·P C1,P C0, T0;XP C10, T0;Xis a Banach space.

Firstly, we consider homogeneous linear impulsive periodic system x. t Axt, t /τk,

Δxt Bkxt, k. 2.2

We introduce the following assumptionH1.

H1.1 A is the infinitesimal generator of a C0-semigroup {Tt, t ≥ 0} on X with domainDA.

H1.2There existsδsuch thatττkT0. H1.3For eachk∈Z0, Bk∈£bXandBBk.

In order to study system2.2, we need to consider the associated Cauchy problem x. t Axt, t∈0, T0\D,

Δxτk Bkk, k1,2, . . . , δ, x0 x.

2.3

(4)

IfxDAandDAis an invariant subspace ofBk, using Theorem 5.2.2,see29, page 144, step by step, one can verify that the Cauchy problem2.3has a unique classical solutionxP C10, T0;Xrepresented byxt St,0x,where

S·,·:Δ t, θ∈0, T0×0, T0|0≤θtT0

−→£bX 2.4

given by

St, θ

⎧⎪

⎪⎪

⎪⎪

⎪⎪

⎪⎪

⎪⎪

⎪⎪

⎪⎪

⎪⎪

⎪⎩

Tt−θ, τk−1θtτk,

T

tτk IBk

T τkθ

, τk−1θ < τk< tτk1, T

tτk

θ<τj<t

IBj T

τjτj−1 IBi

T τiθ

,

τi−1θ < τi≤ · · ·< τk< tτk1. 2.5 The operator {St, θ,t, θ ∈ Δ} is called impulsive evolution operator associated with {Bk;τk}k1.

Now we introduce theP C-mild solution of Cauchy problem2.3andT0-periodicP C- mild solution of the system2.2.

Definition 2.1. For everyxX, the functionxP C0, T0;Xgiven byxt St,0xis said to be theP C-mild solution of the Cauchy problem2.3.

Definition 2.2. A functionxP C0,∞;Xis said to be aT0-periodicP C-mild solution of system2.2if it is aP C-mild solution of Cauchy problem2.3corresponding to somexand xtT0 xtfort≥0.

The following lemma gives the properties of the impulsive evolution operator {St, θ,t, θ∈Δ}associated with{Bk;τk}k1which are widely used in sequel.

Lemma 2.3 see 24, Lemma 1. Impulsive evolution operator {St, θ,t, θ ∈ Δ} has the following properties.

1For 0θtT0,St, θ£bX, that is, sup0≤θ≤t≤T0St, θ ≤MT0, whereMT0 >0.

2For 0θ < r < tT0,r /τk,St, θ St, rSr, θ.

3For 0θtT0andNZ0,StNT0, θNT0 St, θ.

4For 0tT0andNZ0,SNT0t,0 St,0ST0,0N.

5If{Tt, t≥0}is a compact semigroup inX, thenSt, θis a compact operator for 0θ <

tT0.

Here, we note that system2.2has a T0-periodicP C-mild solution xif and only if ST0,0has a fixed point. The impulsive evolution operator{St, θ,t, θ∈Δ}can be used to reduce the existence ofT0-periodicP C-mild solutions for linear impulsive periodic system to the existence of fixed points for an operator equation. This implies that we can build up

(5)

the new framework to study the periodicP C-mild solutions for integrodifferential impulsive periodic system on Banach space.

Consider nonhomogeneous linear impulsive periodic system

xt ˙ Axt ft, t /τk,

Δxt Bkxt ck, k, 2.6

and the associated Cauchy problem

xt ˙ Axt ft, t∈0, T0\D, Δxτk Bkk ck, k1,2, . . . , δ,

x0 x.

2.7

wherefL10, T0;X,ftT0 ftandcck.

Now we introduce theP C-mild solution of Cauchy problem2.7andT0-periodicP C- mild solution of system2.6.

Definition 2.4. A functionxP C0, T0;X, for finite interval0, T0, is said to be aP C-mild solution of the Cauchy problem 2.6 corresponding to the initial value xX and input fL10, T0;Xifxis given by

xt St,0x t

0

St, θfθdθ

0≤τk<t

S t, τk

ck. 2.8

Definition 2.5. A functionxP C0,∞;Xis said to be aT0-periodicP C-mild solution of system2.6if it is aP C-mild solution of Cauchy problem2.7corresponding to somexand xtT0 xtfort≥0.

3. The generalized Gronwall inequality

In order to use Leray-Schauder theorem to show the existence of periodic solutions, we need a new generalized Gronwall inequality with impulse, mixed-type integral operator, andB- norm which is much different from classical Gronwall inequality and can be used in other problemssuch as discussion on integrodifferential equation of mixed type, see15. It will play an essential role in the study of nonlinear problems on infinite-dimensional spaces.

We first introduce the following generalized Gronwall inequality with impulse and B-norm.

Lemma 3.1. LetxP C0,∞, Xand satisfy the following inequality:

xt ≤ab t

0

λ1dθd t

0

xθλB3dθ, 3.1

(6)

wherea, b, d0, 0λ1, λ31 are constants, andxθBsup0≤ξ≤θxξ. Then

xt≤a1ebct. 3.2 Proof. iFor 0≤λ1,λ3<1, letλmax{λ1, λ3} ∈0,1and

yt

⎧⎨

1, xt ≤1,

xt, xt>1.

3.3

Then

xtyt≤a1 b t

0

λdθd t

0

yθλ

B ∀t∈ 0, T0

. 3.4

Using3.4, we obtain ytλ

B≤a1 bd

t

0

yθλ

Bdθ. 3.5

Define

ut a1 bd

t

0

yθλ

Bdθ, 3.6

we get

ut b˙ dytλ

B, t /τk,

u0 a1, u

τk0 u

τk

.

3.7

SinceytλBut, we then have

ut˙ ≤bdut, t /τk,

u0 a1, u

τk0 u

τk

. 3.8

Fort∈τk, τk1, by3.8, we obtain utu

τk0

ebdt−τku τk

ebdt−τk, 3.9

(7)

further,

ut≤a1ebdt, 3.10

thus,

xtyt

B≤a1ebdt. 3.11

iiForλ1λ3 1, we only need to define

u1t a bd t

0

xθ

Bdθ, 3.12

Similar to the proof ini, one can obtain xtxt

Baebdt. 3.13

Combiningiandii, one can complete the proof.

Using Gronwall’s inequality with impulse andB-norm, we can obtain the following new generalized Gronwall Lemma.

Lemma 3.2. LetxP C0, T0;Xsatisfy the following inequality:

xt

ab t

0

λ1dθc T0

0

λ2dθd t

0

xθλ3

Bdθe T0

0

xθλ4

B ∀t∈

0, T0

, 3.14 where λ1, λ3 ∈ 0,1,λ2, λ4 ∈ 0,1,a, b, c, d, e0 are constants. Then there exists a constant M>0 such that

xtM. 3.15

Proof. ByLemma 3.1, we obtain that

xtytyt

Bebdt

a1 c T0

0

λdθe T0

0

yθλ

B

, 3.16

(8)

where

yt

⎧⎨

1, xt≤1, xt, xt>1, λ

⎧⎨

max λ1, λ2, λ3, λ4

∈0,1, if λ1, λ2, λ3, λ4∈0,1, max λ2, λ4

∈0,1, if λ1λ31, λ2, λ4∈0,1.

3.17

Define qt

ebdT0

a1 c t

0

λdθc T0

0

λdθe t

0

yθλ

Bdθe T0

0

yθλ

B

, 3.18

thenqis a monotone increasing function and qt˙

ebdT0

cytλeytλ

B

≤ceebdT0ytλytλ

B

≤2ceebdT0qλt.

3.19

Consider d

dtq1−λt 1−λq−λtqt˙ ≤2ceebdT01−λ. 3.20 Integrating from 0 tot, we obtain

q1−λt−q1−λ0≤2ceebdT01−λt, 3.21 that is,

qt

q1−λ0 2ceebdT01−λt1/1−λ

. 3.22

On the other hand,

2q0 2ebdT0

a1 c T0

0

λdθe T0

0

yθλ

B

;

qT0 ebdT0

a1 2c T0

0

λ2e T0

0

yθλ

B

.

3.23

(9)

Now, we observe that

2q0−ebdT0a1 q T0

q1−λ0 2ceebdT0T01−λ1/1−λ

. 3.24

As a result, we get

2q0−ebdT0a11−λ

q1−λ0≤2ceebdT0T01−λ. 3.25

Letting

Υz

2z−ebdT0a11−λz1−λ−2ceebdT0T01−λ, 3.26

we haveΥ∈CebdT0a1/2,∞;RandΥebdT0a1/2<0. Moreover,

z→lim

Υz

z1−λ21−λ−1>0. 3.27

Hence, there exists enough largez0 > ebdT0a1/2 > 0 such thatΥz > 0 for arbitrary zz0. Meanwhile,Υq0≤0. Thus,q0z0.

As a result, we obtain

xtytq T0

2q0−ebdT0a1

≤2z0ebdT0a1≡M>0 ∀t∈ 0, T0

.

3.28

4. Periodic solutions of integrodifferential impulsive periodic system

In this section, we consider the following integrodifferential impulsive periodic system:

xt ˙ Axt f

t, x, t

0

gt, s, xds

, t /τk, Δxt Bkxt ck, k.

4.1

and the associated Cauchy problem

xt ˙ Axt f

t, x, t

0

gt, s, xds

, t∈0, T0\D, Δx

τk Bkx

τk

ck, k1,2, . . . , δ, x0 x.

4.2

(10)

By virtue of the expression of theP C-mild solution of the Cauchy problem2.7, we can introduce theP C-mild solution of the Cauchy problem4.2.

Definition 4.1. A functionxP C0, T0;Xis said to be aP C-mild solution of the Cauchy problem4.2corresponding to the initial valuexX if xsatisfies the following integral equation:

xt St,0x t

0

St, θf

θ, xθ, θ

0

g

θ, s, xs ds

0≤τk<t

S t, τk

ck fort∈ 0, T0

.

4.3

Now, we introduce theT0-periodicP C-mild solution of system4.1.

Definition 4.2. A functionxP C0,∞;Xis said to be aT0-periodicP C-mild solution of system4.1if it is aP C-mild solution of Cauchy problem4.2corresponding to somexand xtT0 xtfort≥0.

AssumptionH2includes the following.

H2.1f:0,∞×X×XXsatisfies the following.

iFor eachx, y∈X×X,tft, x, yis measurable.

iiFor eachρ >0,there existsLfρ>0 such that, for almost allt∈0,∞and allx1, x2,y1, y2X,x1,x2,y1,y2ρ, we have

f t, x1, y1

f

t, x2, y2Lfρx1x2y1y2. 4.4

H2.2There exists a positive constantMf such that ft, x, yMf

1xy

∀x, y∈X. 4.5

H2.3ft, x, yisT0-periodic int, that is,ftT0, x, y ft, x, y, t≥0.

H2.4LetD {t, s ∈ 0∞×0∞; 0 ≤ st}. The functiong : D×XX is continuous for eachρ > 0, there existsLgρ > 0 such that, for eacht, s ∈ Dand each x, yXwithx,y ≤ρ, we have

gt, s, xgt, s, y≤Lgρx−y. 4.6

H2.5There exists a positive constantMgsuch that gt, s, xMg

1x

∀x, y∈X. 4.7

(11)

H2.6gt, s, xare T0-periodic intands, that is,gtT0, sT0, x gt, s, x, ts≥0 and

T0

0

gt, s, xds0, ts≥0. 4.8

H2.7For eachk∈Z0 andckX, there existsδ∈Nsuch thatcck. Lemma 4.3. Under assumptions [H2.4] and [H2.5], one has the following properties:

1·

0g·, s, xsds:P C0, T0;XP C0, T0;X.

2For allx1, x2P C0, T0;Xandx1P C0,T0;X,x2P C0,T0;Xρ, t

0

g

t, s, x1s ds

t

0

g

t, s, x2s ds

LgρT0x1x2

B. 4.9

3ForxP C0, T0;X, t

0

g

t, s, xs ds

MgT0

1xt

B

. 4.10

Proof. 1Sincegis continuous in its variables and satisfies linear growth conditions, one can verify that·

0g·, s, xsdsmapsP C0, T0;XtoP C0, T0;X.

2Letx1, x2P C0, T0;X,x1P C0,T0;X,x2P C0,T0;Xρ, we have t

0

g

t, s, x1s ds

t

0

g

t, s, x2s ds

t

0

g

t, s, x1s

g

t, s, x2sds

t

0

Lgρx1s−x2sds

Lgρtx1t−x2t

B

LgρT0x1t−x2t

B.

4.11

3ForxP C0, T0;X, t

0

g

t, s, xs ds

Mg

t

0

1xsds

Mgt

1xt

B

MgT0

1xt

B

.

4.12

Now we present the existence ofP C-mild solution for system4.2.

(12)

Theorem 4.4. Assumptions [H1.1], [H2.1], [H2.4], and [H2.5] hold. Then system4.2has a unique P C-mild solution given by

x t, x

St,0x t

0

St, θf

θ, xθ, θ

0

g

θ, s, xs ds

0≤τk<t

S t, τk

ck. 4.13

Proof. A similar result is given by Wei et al.15. Thus, we only sketch the proof here. In order to make the process clear, we divide it into three steps.

Step 1. We consider the following general integrodifferential equation without impulse

xt ˙ Axt f

t, x, t

0

gt, s, xds

, t∈s, τ, xs xX.

4.14

In order to obtain the local existence of mild solution for system4.14, we only need to set up the framework for use of the contraction mapping theorem. Consider the ball given by

B xC s, t1

;X

|xtx≤1, s≤tt1

, 4.15

wheret1would be chosen, andxt ≤1x ρ,stt1.B ⊆Cs, t1, Xis a closed convex set. Define a mapQ onBgiven by

Qxt Ttx t

s

Ttθf

θ, xθ, θ

0

g

θ, s, xs ds

dθ. 4.16

Under assumptionsH1.1,H2.1,H2.2,H2.4,H2.5andLemma 3.1, one can verify that mapQ is a contraction map onBwith chosent1 > 0. This means that system 4.14has a unique mild solutionxCs, t1;Xgiven by

xt Ttx t

s

Tt−θf

θ, xθ, θ

0

g

θ, s, xs ds

on s, t1

. 4.17

(13)

Again, usingLemma 3.1, we can obtain the a priori estimate of the mild solutions for system 4.14and present the global existence of mild solutions.

Step 2. Fort∈τk, τk1, consider the Cauchy problem

xt ˙ Axt f

t, x, t

0

gt, s, xds

, t

τk, τk1 , x

τk

xkIBk

x τk

ckX.

4.18

ByStep 1, Cauchy problem4.18also has a uniqueP C-mild solution

xt T tτk

xk t

τk

Tt−θf

θ, xθ, θ

0

g

θ, s, xs ds

dθ. 4.19

Step 3. Combining all of the solutions onτk, τk1 k1, . . . , δ, one can obtain theP C-mild solution of Cauchy problem4.2given by

x t, x

St,0x t

0

St, θf

θ, xθ, θ

0

g

θ, s, xs ds

0≤τk<t

S t, τk

ck. 4.20

This completes the proof.

To establish the periodic solutions for system4.1, we define a Poincar´e operator from XtoXas follows:

P x

x T0, x S

T0,0 x

T0

0

S T0, θ

f

θ, x θ, x

, θ

0

g θ, s, x

s, x ds

0≤τk<T0

S T0, τk

ck, 4.21

where x·, x denote the P C-mild solution of Cauchy problem 4.2 corresponding to the initial valuex0 x, then we examine whetherPhas a fixed point.

We first note that a fixed point ofPgives rise to a periodic solution.

Lemma 4.5. System4.1has aT0-periodicP C-mild solution if and only ifPhas a fixed point.

Proof. Suppose x·T0, thenx0 xT0 Px0. This implies that x0 is a fixed point of P. On the other hand, if P x0 x0, x0X, then for the P C-mild solution x·, x0of Cauchy problem4.2corresponding to the initial valuex0 x0, we can define

(14)

y· x·T0, x0, theny0 xT0, x0 P x0 x0. Now, fort >0, we can use2,3, and 4ofLemma 2.3and assumptionsH1.2,H1.3,H2.3,H2.6, andH2.7to arrive at

yt x

tT0, x0

S

tT0, T0 S

T0,0 x0

T0

0

S

tT0, T0

S T0, θ

f

θ, x θ, x0

, θ

0

g θ, s, x

s, x0

ds

0≤τk<T0

S

tT0, T0

S T0, τk

ck

tT0

T0

S

tT0, θ f

θ, x

θ, x0

, θ

0

g θ, s, x

s, x0

ds

T0≤τ<tT0

S

tT0, τ c

St,0

S T0,0

x0 T0

0

S T0, θ

f

θ, x θ, x0

, θ

0

g θ, s, x

s, x0

ds

0≤τk<T0

S T0, τk

c

t

0

S

tT0, θT0 f

θT0, x

θT0, x0 ,

θT0

0

g

θT0, s, x s, x0

ds

T0≤τ<tT0

S

tT0, τ c

St,0x T0

t

0

S

tT0, θT0 f

θT0, x

θT0, x0 ,

θT0

T0

g

θT0, s, x s, x0

ds

T0≤τ<tT0

S

tT0, τ c

St,0x T0

t

0

St, θf

θ, x

θT0, x0 ,

θ

0

g

θT0, sT0, x

sT0, x0 ds

T0≤τ<tT0

S

tT0, τ c

St,0y T0

t

0

St, θf

θ, y θ, y0

, θ

0

g θ, s, y

s, y0 ds

0≤τk<t

S t, τk

ck

St,0y0 t

0

St, θf

θ, y θ, y0

, s

0

g θ, s, y

s, y0 ds

0≤τk<t

S t, τk

ck. 4.22

(15)

This implies thaty·, y0is aP C-mild solution of Cauchy problem4.2with initial value y0 x0. Thus the uniqueness implies thatx·, x0 y·, y0 x·T0, x0, so thatx·, x0 is aT0-periodic.

Next, we show thatPdefined by4.21is a continuous and compact operator.

Lemma 4.6. Suppose that {Tt, t ≥ 0} is a compact semigroup in X. Then the operatorP is a continuous and compact operator.

Proof. 1Show thatPis a continuous operator onX.

Letx, y ∈ Ξ⊂ X, whereΞis a bounded subset ofX. Suppose thatx·, xandx·, y are theP C-mild solutions of Cauchy problem4.2corresponding to the initial values xand yX,respectively, given by

x t, x

St,0x t

0

St, θf

θ, x θ, x

, θ

0

g θ, s, x

s, x ds

0≤τk<t

S T0, τk

ck;

x t, y

St,0y t

0

St, θf

θ, x θ, y

, θ

0

g θ, s, x

s, y ds

0≤τk<t

S T0, τk

ck. 4.23

Thus, we obtain x

t, x

MT0x

1MgT0

MT0MfT0MT0

0≤τk<T0

ckMT0Mf

t

0

x

θ, xdθ MT0MfMgT0

θ

0

x

s, xds

a0MT0Mf

t

0

x

θ, xdθMT0MfMgT0

t

0

xs,x

Bds, x

t, y

MT0y

1MgT0

MT0MfT0MT0

0≤τk<T0

ckMT0Mf t

0

x

θ, ydθ MT0MfMgT0

t

0

x

s, yds

b0MT0Mf

t

0

x

θ, ydθMT0MfMgT0

t

0

xs,y

Bds,

4.24

wherexs,xBsup0≤ξ≤sxξ, xandxs,yBsup0≤ξ≤sxξ, y.

ByLemma 3.1, one can verify that there exist constantsM1andM2>0 such that x

t, xM1, x

t, yM2. 4.25

(16)

Let ρ max{M1, M2} > 0, then x·, x,x·, y ≤ ρ which imply that they are locally bounded.

By assumptionsH2.1,H2.2,H2.4,H2.5, and2ofLemma 4.3, we obtain x

t, x

x

t, ySt,0xy

t

0

St, θ f

θ, x

θ, x ,

θ

0

g θ, s, x

s, x ds

f

θ, x θ, y

, θ

0

g θ, s, x

s, y

ds

MT0xyMT0Lfρ t

0

x θ, x

x

θ, ydθ MT0LfρLgρT0

t

0

xs,xxs,y

Bds.

4.26

ByLemma 3.1again, one can verify that there exists constantM3>0 such that x

t, x

x

t, yM3MT0xyLxy, ∀t∈ 0, T0

, 4.27

which implies that P

x

P

yx T0, x

x

T0, yLxy. 4.28 Hence,Pis a continuous operator onX.

2Verify thatPtakes a bounded set into a precompact set inX.

LetΓis a bounded subset ofX. DefineKPΓ {Px∈X|x∈Γ}.

For 0< εT0, define

KεPεΓ S

T0, T0ε x

T0ε, x

|x∈Γ

. 4.29

Next, we show thatKεis precompact inX. In fact, forx∈Γfixed, we have x

T0ε, x

S

T0ε,0 x

T0−ε

0

S

T0ε, θ f

θ, x

θ, x ,

θ

0

g θ, s, x

s, x

ds

0≤τk<T0−ε

S

T0ε, τk ck

(17)

MT0xMT0MfT0

1MgT0

MT0Mf

T0

0

x

θ, xdθ MT0

0≤τk<T0

ckMT0MfMgT0

T0

0

xs,x

Bds

MT0xMT0MfT0

1MgT0

1MgT0

MT0MfT0ρMT0

0≤τk<T0

ck. 4.30 This implies that the set{xT0ε, x|x∈Γ}is totally bounded.

By virtue of {Tt, t ≥ 0} which is a compact semigroup and 5 of Lemma 2.3, ST0, T0εis a compact operator. Thus,Kεis precompact inX.

On the other hand, for arbitraryx∈Γ,

Pε x

S T0,0

x T0−ε

0

S T0, θ

f

θ, x θ, x

, θ

0

g θ, s, x

s, x ds

0≤τk<T0−ε

S T0, τk

ck. 4.31

Thus, having this combined with4.21, we have Pε

x

P x

T0−ε

0

S T0, θ

f

θ, x θ, x

, θ

0

g θ, s, x

s, x ds

T0

0

S T0, θ

f

θ, x θ, x

, θ

0

g θ, s, x

s, x ds

0≤τk<T0−ε

S T0, τk

ck

0≤τk<T0

S T0, τk

ck

T0

T0−ε

S

T0, θ f

θ, x

θ, x ,

θ

0

g θ, s, x

s, x

ds

dθMT0

T0−ε≤τk<T0

ck

MT0Mf

1MgT01ρεMT0

T0−ε≤τk<T0

ck.

4.32

It is shown that the set K can be approximated to an arbitrary degree of accuracy by a precompact setKε. HenceK itself is a precompact set inX. That is,P takes a bounded set into a precompact set inX. As a result,Pis a compact operator.

In order to use Leray-Schauder fixed pointed theorem to examine whether the operator P has a fixed point, we have to make assumptions H2.2 and H2.5 a little stronger as follows.

(18)

H2.2There exists constantNf >0 and 0< λ <1 such that ft, x, yNf

1xλyλ

∀x, y∈X. 4.33

H2.5There exists a positive constantNg>0 and 0< λ <1 such that gt, s, x≤Ng

1xλ

∀x∈X. 4.34

Now, we can give the main results in this paper.

Theorem 4.7. Assumptions [H1], [H2.1], [H2.2], [H2.3], [H2.4], [H2.5], [H2.6], and [H2.7] hold.

Suppose that{Tt, t≥0}is a compact semigroup inX. Then system4.1has aT0-periodicP C-mild solution on0,∞.

Proof. By virtue of{Tt, t≥0}which is a compact semigroup and5ofLemma 2.3,ST0,0 is a compact operator on infinite-dimensional spaceX. Thus,ST0,0/αI,α∈R. Then, there existsβ >0 such thatσST0,0−Ix ≥βxforσ∈0,1. In fact, defineΠσIσST0,0, σ ∈ 0,1, and Πσ : 0,1→£bX and Πσ : 0,1→R. It is obvious thathC0,1;R. Thus, there existσ∈0,1andβ >0 such that

min |σ∈0,1

β >0. 4.35

If not, there exists σ ∈ 0,1such that 0. We can assert thatσ /0 unless 1.

Thus, forσ∈0,1,

S T0,0

1

σI, where 1

σ ≥1, 4.36

which is a contradiction withST0,0/αI,α∈R.

ByTheorem 4.4, for fixed xX, Cauchy problem4.2corresponding to the initial value x0 xhas theP C-mild solution x·, x. ByLemma 4.6, the operatorP defined by 4.21, is compact.

According to Leray-Schauder fixed point theory, it suffices to show that the set{x ∈ X|xσP x, σ ∈0,1}is a bounded subset ofX. In fact, letx∈ {x∈X|xσP x, σ ∈0,1}, we have

βxσS T0,0

I x σ

T0

0

S

T0, θ f

θ, x

θ, x ,

θ

0

g θ, s, x

s, x

ds

σ

0≤τk<T0

S

T0, τkck.

4.37

参照

関連したドキュメント

By applying the method of 10, 11 and using the way of real and complex analysis, the main objective of this paper is to give a new Hilbert-type integral inequality in the whole

By applying the Schauder fixed point theorem, we show existence of the solutions to the suitable approximate problem and then obtain the solutions of the considered periodic

We find the criteria for the solvability of the operator equation AX − XB = C, where A, B , and C are unbounded operators, and use the result to show existence and regularity

Tang, “Explicit periodic wave solutions and their bifurcations for generalized Camassa- Holm equation,” International Journal of Bifurcation and Chaos in Applied Sciences

Abstract: This paper generalizes a Tatar’s result of an impulsive nonlinear singular Gronwall-Bihari inequality with delay [J... Gronwall-Bihari Inequality Shengfu Deng and

The properties of limit periodic homoge- neous linear difference systems with respect to their almost periodic solutions are mentioned, e.g., in [9, 24].. This paper is divided

As an application, we present in section 4 a new result of existence of periodic solutions to such FDI that is a continuation of our recent work on periodic solutions for

We obtain global weighted Caccioppoli-type and Poincar´e inequalities in terms of Orlicz norms for solutions to the nonhomogeneous A-harmonic equation d Ax, dω Bx,