BOUNDED ORDER

THE UNIFORM BOUNDEDNESS PRINCIPLE FOR ORDER BOUNDED OPERATORS

CHARLES SWARTZ

Department of Mathematical Sciences New Mexico State University

Las

Cruces, NM 88003

(Received

_{April 18,}

1988)

ABSTRACT: Under appropriate hypotheses on the spaces, it is shown that a sequence of order bounded llnear operators which is poJntwJse order bounded is uniformly order bounded on order bounded subsets. This result is used to establish a Banach-Stelnhaus Theorem for order bounded operators.

KEY

WORDS AND PHRASES: lattice, order bounded operator, uniform boundedness principle, Banach-Steinhaus Theorem.

1980 NATHEKATICS SUBJECT CLASSIFICATION: 48A40.

1. INTRODUCTION

In

this note we consider the problem of obtaining a version of the classical Uniform Boundedness PrlncJple of functional analysis for linear operators between vector lattices. If X is a Banach space and

Y

is a normed linear space, the Uniform Boundedness Principle then asserts that any sequence

(Tj)

of continuous linear operators from

X

into

Y

which is poJntwise bounded

on X

is such that the sequence of operator norms

(,Till)

^{is bounded}

^{{[1] t4).}

^{It is easy} to see that the condition that the

(llTi.)

are bounded is equivalent to the condition that the sequence

(T i) ^Js

^{uniformly bounded} on bounded subsets of

X

{see the discussion in

[1] 14).

Thus, a possible version of the Uniform Boundedness Principle for order bounded linear operators

(Ti)

^{between vector lattices X and Y might be that if}

(Tj)

^{is pointwise} order bounded on X, then

(Ti}

^{is uniformly} order bounded on order bounded subsets of X. We will show below in Example that such d

straightforward analogue of the Uniform Boundedness Principle does not hold even Jf both X and

Y

are Banach lattices. However, by Imposing additional conditions on

the spaces and by employing the matrix methods of

[1],

we will obtain an order version of the Uniform Boundedness Principle in Theorem 3 below. Using the Uniform Boundedness Principle, we also establish a version of the Banach-Stetnhaus Theorem for order bounded operators which generalizes a result of Nakano.

2. RESULTS

First, conslder ^the following example which shows that a straightforward analogue of the Uniform Boundedness Principle does not ^ho]d for order bounded linear operators between Banach lattices.

(A

linear operator T between vector lattices or Riesz spaces X and is order bounded If T carries order bounded subsets of X into order bounded subsets of

,

where a subset A of a vector lattice X order bounded if there is an order Interval

[-u, u] (v X

-u v

u)

such that A^c

[-u, u]

(in genera], we conform to the notation and terminology of

[5]).)

EXAMPLE

1. Let X

LI[O,1]

^{and Y} c

O ^and assume that these spaces have the usual polntwise ordering. For f e

L1[0,1],

set

fk

J

/

^{f(t)sin ktdt. Define}

0 k

Tk

X Y

by

Tkf ^fjej,

^where

^ej

^{is the element of}

^{c o}

^{with a in the}

J=l

jth

coordinate and 0 in the other coordinates. Each T

k ^is order bounded since k

if

Ifl ^<

^{g in X, then}

If j[ og

^{a and}

^ITk

^f

^{< aej AI}

^{so the sequence}

(Tk)

^{is pointwise order bounded on X since if f e X, then}

[Tkf [fjJej

^{e c}0

(note ([fj[)

^{e c}

o

^{by the} Rlemann-Lebesgue

Lemma).

j=l

However,

(Tk}

^{is not uniformly} order bounded on order bounded subsets of X since if

Vk(t)

^{sin kt, then}

(Fk)

^{is order bounded in}

^X (IFkl ^I)

^but

(Tl(i)} (el/2)

^{is not} order bounded in

c

O

Note

that both X and Y in this example are Banach lattices and both are Dedekind complete.

In

order to obtain our version of the Uniform Boundedness Principle for ordered spaces, we first obtain a matrix theorem which is the analogue for ordered spaces of the matrix result given in

[I]

2.1.

Throughout the remainder of this note we let

X

and Y denote Riesz spaces

(vector

lattices), h sequence

(x k)

^{in X}

^Is

u-convergent to

x,

where u O, if there exists a scalar sequence

tk40

^{such that}

IXk

^x

^tku;

^{we write}

u lim

x

k x. The element u is called a convergence ^regulator for

(Xk).

^The

sequence

{x k)

^is

^reltivel

^niformly convergent to

x

if

(x k)

^is u-convergent to

x

for some convergence regulator u O; we write r llm x

k x.

LEMMA 2.

([4]

2.1) Let the infinite matrix

[Xlj], xij

^{X, be such that its}

rows and columns are u-convergent to O. Let

lJ

^{> O,}

6iJ "

^Then there exists an increasing sequence of positive integers

(pi)

^{such that}

IXpiPj ^iju

^for

PROOF. Put

Pl

^{1. Since}

(xlj)

^and

{Xil)

are u-convergent to ^O, there exists

P2

^>

Pl

^{such that}

IXlp21 ^{_< 12u}

^and

IXp21 ^{< 21u"}

^{Similarly, there}

iS

P3 ^> P2

^{such that}

IXplP31 ^el3U, IXp2P31 ²³

^u,

IXp3P21 ³²

^{u and}

IXp3pl ^e31u.

^{Now continue.}

We now prove our Uniform Boundedness Principle for order bounded operators.

Recall that if Y is Dedekind complete, then a linear map T X

Y

is order bounded if and only if

T

is regular

([5] VIII.2.2).

THEOREM 3. Let g be Dedekind o-complete and let

Y

be Dedeklnd complete and have an order unit u. If

T

1 X Y is a sequence of order bounded linear maps which is pointwise order bounded on X, then

(T i)

is uniformly order bounded on order bounded subsets of g.

PROOF. If the conclusion fails, there is

an

interval

I-w, w]

in

X

such that

(Tt([-w, ^w])

^e

⁾

^is not order bounded. Thus, for each there exist

x

_i ^e

[-w, w]

and m

i such that

TmtX

^l

^{14j-u, u]. For}

^notational convenience, assume that m

I

^{i. Then}

14j-u, u]. (2.1)

Tlx

1

Now consider the matrix M

[(1/t)Tl(xj/J2)].

^{For each}

^J, (Tl(xj/j2))

1 ^is order bounded

so

the

jth

column of

M

is relatively uniformly convergent to O. The sequence

(xj/J 2)

is relatively uniformly convergent to 0, and since each T

t

^is sequentially continuous with

respect

to relative uniform convergence

([5]

VIII

.2),

the

I

^th row of

M

is also relatively uniformly convergent to 0. Thus, the rows and columns of M are u-convergent to O. By Lemma 2 there is an increasing sequence of positive integers

(pi)

^{such that}

I(1/pi)Tpi(Xpj/pj2) ^2-l-Jr

^for

#

J.

^Again for notational convenience, we assume that

Pi

^{i. Since}

^X

^is

-complete and

Ixjl

^{w, the series}

j/j2

^is absolutely order convergent to an element x

X ([5]

IV. 9); moreover, this series is actually w-convergent in X

n

since

Ix xj/j21 Ixjl/J2 ( ^1/J2)w.

^{From the continuity of T}

J=l

^j=n+l

Jfn+l

with respect to relative uniform convergence, we have

< +1

.1=1

j=l

.j=l

< l(]/,)T,(x)l

^j=l

.

²

_< (]/J)lTi(x)l

⁺

2-iv ^(2.2)

where we have used the order completeness of Y to insure that the series on the right hand side of

(2)

are convergent. Both terms on the right hand side of

(2)

are order bounded. Hence, there exists k such that

(1/t)lTl(xi/t2) ^{k[-u, u]}

^for

all t. Putting 1 k contradicts

(1)

and the result

Is

established.

Note that the range space c

O in Example is Dedeklnd complete, but does not have an order unit.

It is perhaps worthwhile noting that if

(T1)

^is uniformly order bounded on order bounded subsets of X, then the sequence of modull

(ITII)

^{also has this}

property. For if 0 x u and

ITt[0, ^u]

^{w, then}

ITllx ^sup(lTizl

⁰

^z

^$

^x)

^w

([5] VIII.2)

so

ITil[0, ^u]

From Theorem 3 we

can

also obtain an order analogue of the equicontlnuity conclusion of the classical Uniform Boundedness Principle. Recall that if

(T 1)

^is

a sequence of continuous linear operators from a normed space

X

into a normed space Y, then

(T i)

^is equlcontlnuous if and only if

(TI)

is uniformly bounded on bounded subsets of

X

if and only if

Tix

llm

Tix

⁰ uniformly in 1 whenever

xj

⁰

^([1]

^{4.5). It is easy to establish}

J

order analogues of these equlcontlnuity conditions for operators which satisfy the conclusion of Theorem 3. We give order analogues of these conditions below and consider the relationship between them. If for each a e A,

{xia

^{is a sequence in}

X, we say that

ia

^is u-convergent to 0 uniformly for a A if there exists a sequence

tl0

^{such that}

IXla ttu

^{for all a A; we write u ltm}

Xia

⁰

uniformly for a ^e A.

PROPOSITION 4. Let T X Y be order bounded. Consider 1

(i)

(T1)

^is uniformly order bounded on order bounded subsets of

X.

(11)

if

r

llm

xj

^{O, then v llm}

Tlx

J

0 uniformly for e for some v e Y, v

>

O,

(111)

r 11m

Tlx

^{0 whenever r Jim x 0.}

Always

(I)

implies

(II)

implies (ill). If

Y

is Archlmedean and has the boundedness property

([3]

1.5.12), then (iii) implies

{I).

PROOF. Assume (i) holds. To establish {ii), there exists

tj,

^{such that}

{tjxj}

^is relatively uniformly convergent to 0

([4]

VI.4). Since

(tjxj)

^is

order bounded, by (i) there is v e Y such that

ITi{tjxj)l

^{_( v for all I,}

^J.

Then

ITixjl ^<_ (1/tj)v

^Implles

^(I11.

Clearly (li) implies (ill). Assume {Ill) and that

Y

is Archimedean with the boundedness property. Let u

>_

O, u e X. By the boundedness property, it suffices to show that if

(Tkl)

^{is a} subsequence,

Ixi ^<_

^{u and}

^tl0,

^then

r lim

tlTklX ^I

^{O. But, since u lim}

^tix

^{i O, this} follows from (iii).

Theorem 3 gives sufficient conditions for the equlcontlnulty condition

(I)

and, therefore,

(ll)

and {lli), to hold.

As in the classical case we can apply the Uniform Boundedness Principle given in Theorem 3 to obtain a Banach-Stelnhaus type result for order bounded operators

([1] 5.1).

There is an order version of the Banach-Steinhaus Theorem for linear functlonals due to Nakano glven in

[5] IX.

1.1, and a general form of the Banach-Steinhaus Theorem for operators between

vector

lattices given in

[4].

^The results in

[4]

treat a different class of operators than that considered below

In

Corollary 5 and the assumptions on the spaces are not as restrictive.

COROLLARY 5. Let X, Y be as in Theorem 3 and let T X-, Y be order

I

bounded. If 0 lim

Tix ^Tx

^{exists for each x e X, then}

^T

^X

^{Y Is}

^{an order}

bounded linear operator.

PROOF.

For

each x,

(TlX)

^{is order bounded since the sequence is order}

convergent. Therefore, from Theorem 3,

(T i)

is uniformly order bounded on order bounded subsets of X. If

[u, v]

is

an

order interval in X, then there is an order interval

I-w, w]

in Y such that

Ti([u, ^{v])c_ [-w, w]}

^{for all i. Hence,}

T{[u, v])

^c

I-w, w]

and

T

is order bounded.

The sequence of operators

(T i)

^{in Example} is pointwlse order convergent to

the operator

T

X

Y

given by Tf

fjej.

^{IIowever, T is not order bounded}

j=l

since

(TFj) (ej/2),

^where

Vj(t)

^{sln//Jt. This shows that in general the}

Banach-Stelnhaus result will not hold if the range space is merely Dedektnd complete.

The author would like to thank Joe

KJst

for his kind assistance.

REFERENCES

I. P. Antoslk and C. Swartz, Matrix Methods in .nalysls, Lecture Notes in Mathematics 1113, Springer-Verlag, 1985.

2. W.A.J. Luxemburg and A.C. Zaanen, Rlesz Spaes I, North Bolland, 1971.

3. A.L. Peressini, Ordered Topological

Vector Soace,

Harper and Row, 1967.

4. C. Swartz, The Banach-Stelnhaus Theorem for ordered spaces, proceedings of the Conference 6FAC, Dubrovnlk, Yugoslavla, 1987, to appear.

5.

B.Z.

Vulikh,

Introduction

^to

the

^Theory

of

^partlallv

Ordered

^Spaces,

Wolters-Noordhoff, 1967.

Special Issue on

Intelligent Computational Methods for Financial Engineering

Call for Papers

As a multidisciplinary field, financial engineering is becom- ing increasingly important in today’s economic and financial world, especially in areas such as portfolio management, as- set valuation and prediction, fraud detection, and credit risk management. For example, in a credit risk context, the re- cently approved Basel II guidelines advise financial institu- tions to build comprehensible credit risk models in order to optimize their capital allocation policy. Computational methods are being intensively studied and applied to im- prove the quality of the financial decisions that need to be made. Until now, computational methods and models are central to the analysis of economic and financial decisions.

However, more and more researchers have found that the financial environment is not ruled by mathematical distribu- tions or statistical models. In such situations, some attempts have also been made to develop financial engineering mod- els using intelligent computing approaches. For example, an artificial neural network (ANN) is a nonparametric estima- tion technique which does not make any distributional as- sumptions regarding the underlying asset. Instead, ANN ap- proach develops a model using sets of unknown parameters and lets the optimization routine seek the best fitting pa- rameters to obtain the desired results. The main aim of this special issue is not to merely illustrate the superior perfor- mance of a new intelligent computational method, but also to demonstrate how it can be used e ff ectively in a financial engineering environment to improve and facilitate financial decision making. In this sense, the submissions should es- pecially address how the results of estimated computational models (e.g., ANN, support vector machines, evolutionary algorithm, and fuzzy models) can be used to develop intelli- gent, easy-to-use, and/or comprehensible computational sys- tems (e.g., decision support systems, agent-based system, and web-based systems)

This special issue will include (but not be limited to) the following topics:

•

Computational methods

: artificial intelligence, neu- ral networks, evolutionary algorithms, fuzzy inference, hybrid learning, ensemble learning, cooperative learn- ing, multiagent learning

•

Application fields

: asset valuation and prediction, as- set allocation and portfolio selection, bankruptcy pre- diction, fraud detection, credit risk management

•

Implementation aspects

: decision support systems, expert systems, information systems, intelligent agents, web service, monitoring, deployment, imple- mentation

Authors should follow the Journal of Applied Mathemat- ics and Decision Sciences manuscript format described at the journal site http://www.hindawi.com/journals/jamds/.

Prospective authors should submit an electronic copy of their complete manuscript through the journal Manuscript Track- ing System at http://mts.hindawi.com/, according to the fol- lowing timetable:

Manuscript Due December 1, 2008 First Round of Reviews March 1, 2009 Publication Date June 1, 2009

Guest Editors

Lean Yu, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100190, China;

Department of Management Sciences, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong;

[email protected]

Shouyang Wang, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100190, China; [email protected]

K. K. Lai, Department of Management Sciences, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong; [email protected]

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