• 検索結果がありません。

SMOOTH BANACH SPACES

N/A
N/A
Protected

Academic year: 2022

シェア "SMOOTH BANACH SPACES"

Copied!
7
0
0

読み込み中.... (全文を見る)

全文

(1)

http://ijmms.hindawi.com

© Hindawi Publishing Corp.

ITERATIVE SOLUTIONS OF K-POSITIVE DEFINITE OPERATOR EQUATIONS IN REAL UNIFORMLY

SMOOTH BANACH SPACES

ZEQING LIU, SHIN MIN KANG, and JEONG SHEOK UME (Received 2 October 2000)

Abstract.Let Xbe a real uniformly smooth Banach space and let T :D(T )⊆X→X be aK-positive definite operator. Under suitable conditions we establish that the iterative method by Bai (1999) converges strongly to the unique solution of the equationT x=f,f∈ X. The results presented in this paper generalize the corresponding results of Bai (1999), Chidume and Aneke (1993), and Chidume and Osilike (1997).

2000 Mathematics Subject Classification. 47H06, 47H07, 47H14.

1. Introduction and preliminaries. LetXbe a real Banach space with a dual space X. The normalized duality mappingJ:X→2X is defined by

J(x)=

f∈X:x, f = x2= f2

, x∈X. (1.1)

It is known thatXis uniformly smooth (equivalently,Xis uniformly convex) if and only ifJis single-valued and uniformly continuous on any bounded subset ofX.

Chidume and Aneke [3] introduced the concept ofK-positive definite operators and established the existence of the unique solution of the equationT x=ffor that oper- ator in real separable Banach spaces. Meanwhile they constructed, inLp(orlp) spaces withp≥2, an iteration method which converges strongly to the unique solution, pro- vided thatT andKcommute. Chidume and Osilike [5] gave a new iteration scheme, in separableq-uniformly smooth Banach spaces, which converges strongly to the unique solution of the equationT x=f,f∈X.

Recently, Bai [1] constructed a more general iteration procedure and improved the results of [3,5] to separable uniformly smooth real Banach spaces.

Very recently, Zhou et al. [7] established the following excellent result, which is a generalization of the main result of Chidume and Aneke [3].

Lemma1.1(see [7]). LetXbe a real Banach space and letT be aK-positive definite operator withD(T )=D(K). Then there exists a constantα >0such that

T x ≤αKx, x∈D(T ). (1.2) Moreover, the operatorT is closed,R(T )=X, and the equationT x=ffor eachf∈X, has a unique solution.

The purpose of this paper is to study the convergence problem of the iteration pro- cedure introduced in [1] forK-positive definite operators in real uniformly smooth real

(2)

Banach spaces. Our results extend the corresponding results due to Bai [1], Chidume and Aneke [3], and Chidume and Osilike [5].

In what follows, we will also need the following concepts and results.

Definition1.2(see [3, 7]). LetX be a real Banach space andX1 a subspace of X. An operatorT with domain D(T )⊇X1 is calledcontinuously X1-invertibleif T, as an operator restricted toX1, has a bounded inverse onR(T ). A linear unbounded operatorT with domainD(T )inXand rangeR(T )inXis calledK-positive definiteif there exist a continuouslyD(T )-invertible closed linear operatorKwithD(A)⊆D(K) and a constantc >0 such that

T u, j(Ku)

≥cKu2, u∈D(T ), j(Ku)∈J(Ku). (1.3) Let X be a real Banach space. Recall that the modulus of smoothness of X is defined by

ρX(t)=sup 1

2

x+y+x−y

1 :x, y∈X, x =1,y ≤t

, t≥0. (1.4) Xis said to beuniformly smoothif limt→0ρX(t)/t=0. Letp >1 be a real number.X is calledp-uniformly smoothif there exists a constantr >0 such that

ρX(t)≤r tp, t >0. (1.5) Hilbert spaces,Lp(orlp) spaces, 1< p <∞, and the Sobolev spacesWpm, 1< p <∞, are allp-uniformly smooth. It is well known that the class ofp-uniformly smooth real Banach spaces is a proper subclass of that of uniformly smooth real ones.

Lemma1.3(see [4,6]). LetXbe a real uniformly smooth Banach space. Then (i) there exist some positive constantsAandBsuch that

x+y2≤ x2+2

y, J(x)

+Amax

x+y, B ρX

y

, x, y∈X. (1.6) (ii) there exists a continuous nondecreasing functionb:[0,∞)→[0,∞)such that

b(0)=0, b(ct)≤cb(t), c≥1;

x+y2≤ x2+2

y, J(x) +max

x,1

yb(y), x, y∈X. (1.7) Lemma1.4(see [2]). Suppose that{αn}n=0,{βn}n=0, and{ωn}n=0are nonnegative sequences such that

αn+1 1−ωn

αnnωn, n≥0, (1.8)

with{ωn}n=0⊂[0,1], n=0ωn= ∞andlimn→∞βn=0. Thenlimn0αn=0.

Lemma1.5(see [6]). LetXbe a real Banach space. Then (i) ρX(0)=0,ρX(t)≤t,t >0;

(ii) ρX(t)is convex, continuous, and nondecreasing on[0,∞);

(iii) ρX(t)/tis nondecreasing on(0,∞).

(3)

2. Main results

Theorem2.1. LetXbe a real uniformly smooth Banach space and letT :D(T )⊆ X→Xbe aK-positive definite operator withD(T )=D(K). Define a sequence{xn}n=0

iteratively from anyf∈Xandx0∈D(T )by

yn=xn+bnvn, xn+1=yn+anun, n≥0; (2.1) vn=K1f−K1T xn, un=K1f−K1T yn, n≥0, (2.2) where{an}n=0and{bn}n=0are arbitrary nonnegative sequences such that

n=0

an+bn

= ∞; (2.3)

n→∞liman=lim

n→∞bn=0; (2.4)

max an, bn

1

2c, n≥0; (2.5)

αAmax

1+αanKv0,

1+αbnKv0, B

2cKv0, n≥0, (2.6) wherec, α, AandBare the constants appearing in (1.2), (1.3), and (1.6), respectively.

Then the sequence{xn}n=0converges strongly to the unique solution of the equation T x=f.

Proof. It follows fromLemma 1.1that the equationT x=fhas a unique solution inX. Note thatT andKare linear. From (2.1) and (2.2) we have

Kvn+1=f−T xn+1=Kun−anT un, n≥0; (2.7) Kun=f−T yn=Kvn−bnT vn, n≥0. (2.8) In view of (2.8) and (1.2), (1.3), and (1.6), we conclude that

Kun2=Kvn−bnT vn2

≤Kvn22bn

T vn, J Kvn

+AmaxKvn+bnT vn, B

ρX

bnT vn

12cbnKvn2 +Amax

1+αbnKvn, B ρX

αbnKvn

(2.9)

for alln≥0. Using (2.7) and (1.2), (1.3), and (1.6), we have Kvn+12=Kun−anT un2

≤Kun22an

T un, J Kun

+AmaxKun+anT un, B

ρX

anT un

12canKun2 +Amax

1+αanKun, B ρX

αanKun

(2.10)

(4)

for alln≥0. SetM= Kv0. We claim that

maxKvn,Kun≤M, n≥0. (2.11)

By virtue of (2.6), (2.9), andLemma 1.5, we get that Ku02

12cb0Kv02 +Amax

1+αb0Kv0, B ρX

αb0Kv0

12cb0

M2+Amax 1+αb0

M, B αb0M

≤M2.

(2.12)

That is, (2.11) is true forn=0. Suppose that (2.11) holds for somen≥0. Using (2.10), (2.6), andLemma 1.5, we infer that

Kvn+12

12canKun2 +Amax

1+αanKun, B ρX

αanKun

1−2can

M2+Amax 1+αan

M, B αanM

≤M2.

(2.13)

From (2.6), (2.9), (2.13), andLemma 1.5, we have Kun+12

12cbn+1Kvn+12 +Amax

1+αbn+1Kvn+1, B ρX

αbn+1Kvn+1

1−2cbn+1

M2+Amax

1+αbn+1 M, B

αbn+1M

≤M2.

(2.14)

Therefore (2.11) holds for all n 0. Since X is uniformly smooth, by (2.4) and Lemma 1.5we conclude that there exist nonnegative sequences{sn}n=0and{tn}n=0

such thatρX(αMan)=snan,ρX(αMbn)=tnbnfor alln≥0 and

n→∞limsn=lim

n→∞tn=0. (2.15)

It follows from (2.5), (2.9), (2.10), and (2.11) that Kvn+12

12can

12cbnKvn2 +

12can

Amax

1+αbnKvn, B ρX

αbnKvn +Amax

1+αanKun, B ρX

αanKun

1−2c

an+bn

+4c2anbnKvn2 +Amax{(1+α)M, B}

ρX αMan

X

αMbn

1−c

an+bnKvn2+L

ansn+bntn

(2.16)

for alln≥0, whereL=Amax{(1+α)M, B}.Let αn=Kvn2, ωn=c

an+bn

, βn=L

crn, n≥0, (2.17)

(5)

where

rn=





0, an+bn=0,

an

an+bn

sn+ bn

an+bn

tn, an+bn=0. (2.18) It follows from (2.15) that limn→∞rn=0. That is, limn→∞βn=0.Thus (2.15) can be rewritten in the form

αn+1 1−ωn

αnnβn, n≥0. (2.19)

Note that (2.3) and (2.5) mean that n=0ωn=∞,ωn∈[0,1]. Consequently,Lemma 1.4 ensures thatαn0 asn→ ∞. That is,

Kvn →0 asn → ∞. (2.20)

It follows from (2.2) and (2.20) that

T xn−f=Kvn →0 asn → ∞. (2.21) Note thatT has a bounded inverse. Thus (2.21) means that xn→T−1f, the unique solution ofT x=f. This completes the proof.

Theorem 2.2. LetX, T, K, f, {xn}n=0, {yn}n=0, {vn}n=0 and {un}n=0 be as in Theorem 2.1. Suppose that {an} and {bn}n=0 are any nonnegative sequences such that (2.3), (2.4), and (2.5) and

max b

αan

, b αbn

2c max

1,Kv0, n≥0, (2.22) whereb(t)is as in (1.7),αandc are the constants appearing in (1.3) and (1.2), re- spectively. Then the sequence{xn}n=0converges strongly to the unique solution of the equationT x=f.

Proof. SetM=max{1,Kv0}.As in the proof of Theorem 3 in [1] we have Kvn+12

1−c

an+bnKvn2+M3α anb

αan

+bnb αbn

, n≥0. (2.23)

Let

αn=Kvn2, ωn=c an+bn

, βn

cM3rn, n≥0, (2.24) where

rn=







0, an+bn=0,

an

an+bn

b αan

+ bn

an+bn

b αbn

, an+bn=0.

(2.25)

It is easily seen that limn→∞βn = 0. The rest of the argument now follows as in the proof of Theorem 2.1 to yield that xn →T−1f as n→ ∞. This completes the proof.

(6)

Remark2.3. Theorems2.1and 2.2extend Theorem 3.3 of Bai [1], Theorem 2 of Chidume and Aneke [3] and Theorem of Chidume and Osilike [5], respectively, in the following ways:

(a) Condition (2.3) is much weaker than n=0an= ∞of [1].

(b)Lp(orlp) spaces,p≥2,in [3] andq-uniformly smooth Banach space,q >1, in [5] are replaced by the more general uniformly smooth Banach spaces.

(c) The commutativity condition ofT andKin [3] is dropped.

(d) The iteration methods in [3,5] are special cases of our iteration method.

Acknowledgements. The first author is grateful to the Research Institute of Nat- ural Sciences, Gyeongsang National University, in Korea for giving him the opportunity to visit the institute, the second author was supported by Korea Research Foundation Grant (KRF-99-005-D00003), and the third author was supported by KOSEF research project No. (2001-1-10100-005-2).

References

[1] C. Bai,Approximation of a solution for aK-positive definite operator equation in uniformly smooth separable Banach spaces, J. Math. Anal. Appl.236(1999), no. 2, 236–242.

MR 2000f:47021. Zbl 946.47007.

[2] S. S. Chang, Y. J. Cho, B. S. Lee, J. S. Jung, and S. M. Kang,Iterative approximations of fixed points and solutions for strongly accretive and strongly pseudo-contractive mappings in Banach spaces, J. Math. Anal. Appl.224(1998), no. 1, 149–165.MR 99g:47146.

Zbl 933.47040.

[3] C. E. Chidume and S. J. Aneke,Existence, uniqueness and approximation of a solution for aK-positive definite operator equation, Appl. Anal. 50 (1993), no. 3-4, 285–294.

MR 95e:47009. Zbl 788.47051.

[4] C. E. Chidume and C. Moore, The solution by iteration of nonlinear equations in uni- formly smooth Banach spaces, J. Math. Anal. Appl.215 (1997), no. 1, 132–146.

MR 98m:47107. Zbl 906.47050.

[5] C. E. Chidume and M. O. Osilike, Approximation of a solution for a K-positive defi- nite operator equation, J. Math. Anal. Appl.210(1997), no. 1, 1–7.MR 98c:47010.

Zbl 901.47002.

[6] Z. B. Xu and G. F. Roach, Characteristic inequalities of uniformly convex and uni- formly smooth Banach spaces, J. Math. Anal. Appl.157 (1991), no. 1, 189–210.

MR 92i:46023. Zbl 757.46034.

[7] H. Y. Zhou, S. M. Kang, and Y. J. Cho,Constructive solvability ofk-positive definite operator equations in uniformly smooth Banach spaces, submitted to Appl. Math. Lett.

Zeqing Liu: Department of Mathematics, Liaoning Normal University, Dalian, Liaon- ing116029, China

E-mail address:[email protected]

Shin Min Kang: Department of Mathematics, Gyeongsang National University, Chinju660-701, Korea

E-mail address:[email protected]

Jeong Sheok Ume: Department of Applied Mathematics, Ghangwon National Univer- sity, Changwon641-773, Korea

E-mail address:[email protected]

(7)

Special Issue on

Boundary Value Problems on Time Scales

Call for Papers

The study of dynamic equations on a time scale goes back to its founder Stefan Hilger (1988), and is a new area of still fairly theoretical exploration in mathematics. Motivating the subject is the notion that dynamic equations on time scales can build bridges between continuous and discrete mathematics; moreover, it often revels the reasons for the discrepancies between two theories.

In recent years, the study of dynamic equations has led to several important applications, for example, in the study of insect population models, neural network, heat transfer, and epidemic models. This special issue will contain new researches and survey articles on Boundary Value Problems on Time Scales. In particular, it will focus on the following topics:

Existence, uniqueness, and multiplicity of solutions

Comparison principles

Variational methods

Mathematical models

Biological and medical applications

Numerical and simulation applications

Before submission authors should carefully read over the journal’s Author Guidelines, which are located at http://www .hindawi.com/journals/ade/guidelines.html. Authors should follow the Advances in Difference Equations manuscript format described at the journal site http://www.hindawi .com/journals/ade/. Articles published in this Special Issue shall be subject to a reduced Article Processing Charge of C200 per article. Prospective authors should submit an elec- tronic copy of their complete manuscript through the journal Manuscript Tracking System at http://mts.hindawi.com/

according to the following timetable:

Manuscript Due April 1, 2009 First Round of Reviews July 1, 2009 Publication Date October 1, 2009

Lead Guest Editor

Alberto Cabada,

Departamento de Análise Matemática, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain; [email protected]

Guest Editor

Victoria Otero-Espinar,

Departamento de Análise Matemática, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain;

[email protected]

Hindawi Publishing Corporation http://www.hindawi.com

参照

関連したドキュメント

After the continuity and compactness of Poincar´e operator P are shown, the existence of T 0 - periodic P C-mild solutions for semilinear impulsive periodic system is established

In order to extend and generalize some recent fixed point results of mixed monotone mappings, Lakshmikantham and ´ Ciri´ c [6] introduced the concept of the mixed g-monotone

The existence of a unique classical nonnegative solution is established and the sufficient conditions for the solution that exists globally or blows up in finite time are

Having established the existence of regular solutions to a small perturbation of the linearized equation for (1.5), we intend to apply a Nash-Moser type iteration procedure in

The existence of periodic solution for a certain functional differential equation with quasibounded nonlinearity is established.. KEY WORDS

In this paper, some results concerning the existence of common fixed points, coincidence points and approxi- mating fixed points with PPF dependence for the pairs of operators in

The starting point for the definition of our notion of orientability for Fredholm maps is a preliminary concept of orientation for Fredholm linear operators of index zero between

The starting point for the definition of our notion of orientability for Fredholm maps is a preliminary concept of orientation for Fredholm linear operators of index zero between