• No results found

Ostrowski type inequalities for sets and functions of bounded variation

N/A
N/A
Protected

Academic year: 2020

Share "Ostrowski type inequalities for sets and functions of bounded variation"

Copied!
16
0
0

Loading.... (view fulltext now)

Full text

(1)

R E S E A R C H

Open Access

Ostrowski type inequalities for sets and

functions of bounded variation

Oleg V Kovalenko

*

*Correspondence: [email protected] Department of Mechanics and Mathematics, O. Honchar Dnipro National University, pr. Gagarina, 72, Dnipro, 49010, Ukraine

Abstract

In this paper we obtain sharp Ostrowski type inequalities for multidimensional sets of bounded variation and multivariate functions of bounded variation.

MSC: Primary 26D15; secondary 41A17; 41A44; 41A63

Keywords: Ostrowski type inequalities; multivariate functions of bounded variation; sets of bounded variation

1 Introduction

In  Ostrowski [] proved the following inequality.

Theorem Let f: [, ]→Rbe a differentiable on(, )function with bounded on(, )

derivative.Then,for all x∈[, ],

f(t)dtf(x)≤

 +

x– 

sup

t∈(,)

f(t).

The inequality is sharp.

Inequalities that estimate deviation of a function from its mean value using different characteristics of the function are usually called Ostrowski type inequalities. Such inequal-ities have many applications, in particular in the area of numerical methods, and are heav-ily studied. See [] and the references therein for results connected with Ostrowski type inequalities for univariate functions of bounded variation and their applications.

The goal of this article is to obtain sharp Ostrowski type inequalities for multivariate functions and multidimensional sets of bounded variations. There are several ways to ex-tend the notion of bounded variation to multivariate functions, see [] for a review of dif-ferent approaches for functions of two variables; [] for the point of view that is generally accepted in literature now.

We introduce a new definition of bounded variation that is based on the Kronrod-Vitushkin approach []. The introduced variation of a multivariate function has (unlike any of the Kronrod-Vitushkin variations) the following two properties: the variation does not change if the argument of the function is multiplied by a non-zero constant; and the

(2)

variation of a multivariate radial function is twice bigger than the variation of the gener-ating one-dimensional function, see Properties  and  below for rigorous statements of the properties.

The paper is organized as follows. In Section  we list the notations used throughout the paper. In Sections  and  we introduce definitions, justify the correctness of the def-initions, and list some properties of the sets and function variations. Section  is devoted to Ostrowski type inequalities.

2 Notations

For a setF⊂Rd, denote byF,intFandFits boundary, interior and closure, respectively.

For arbitraryt∈R, set Rd

t :={(x,t)∈Rd: x∈Rd–}. Forx,y∈Rd, by xywe denote the

segment with ends in the pointsxandy, i.e.,xy={( –t)x+ty:t∈[, ]}. Forc∈Rand F⊂Rd, setcF:={cx:xF}.

For two setsF,F⊂Rd, setρ(F,F) :=infxF,yF|xy|, where|w| denotes the Eu-clidean distance between the pointw∈Rdand zero elementθ ofRd.

Forε≥, two setsF,F⊂Rdare calledε-disjoint ifρ(F,F) >ε. Obviously, two

com-pact setsF,F⊂Rdare -disjoint if and only if they are disjoint. Forε> , a setFis called

ε-connected if there does not exist a partitionF=F∪Fintoε-disjoint non-empty sets

F,F. Some properties ofε-components can be found in Appendix I of [].

For an arbitrary function f: E⊂RdRandcR, denote by{f c} the set {x

E:f(x)≥c}. Similarly, we define the sets{fc}and{f =c}.

Denote bySd–thed–  dimensional unit sphere{xRd:|x|= }; forε>  andxRd,

Bd(x,ε) :={yRd:|xy| ≤ε};Bd(ε) :=Bd(θ,ε) andBd:=Bd().

ByPd–we denote thed–  dimensional real projective space, i.e., the set of all lines

inRdthat containθ. The distance between two linesr

,r∈Pd–is by definition equal

to the angle betweenr andr. The measure of a set A⊂Pd–is by definition equal to

the spherical measure of the setlAlSd–; so that the measure ofPd–is equal to the

measure ofSd–.

For eachr∈Pd–, byd–(r) we denote the hyperplane that containsθand is orthogonal

to the liner;d–(r) is considered as ad– -dimensional space withd– -dimensional

Lebesgue measure and Euclidean metric. For eachβd–(r), byl(r,β) we denote the

line that containsβand is parallel tor.

It is assumed that product topology is induced on the Cartesian product of a finite num-ber of topological spaces and product measure is induced on the Cartesian product of a finite number of measure spaces. By μk,k∈N, we denote thek-dimensional Lebesgue measure inRk; byμwe denote the spherical measure on the unit sphereSd–and the measure on the projective spacePd–.

3 Variation of a set

3.1 Definition

Definition  Denote byN(F) the number of connected components of the setF⊂Rd; 

for an empty set, and +∞if the set of connected components is infinite.

(3)

Definition  For a compact setF⊂Rdand a linerPd–, set

v(F,r) := ess sup βd–(r)

NFl(r,β).

Definition  For a compact setF⊂Rdand a numberp∈[,∞], set

vp(F) :=

⎧ ⎨ ⎩

(μPd–

Pd–vp(F,r)dr)

p, p[,),

ess suprPd–v(F,r), p=∞.

Remark  Ifd= , then for allp∈[,∞] we setvp(F) =N(F).

Definition  Let a compact setF⊂Rd,ε andp[,] be given. Set

Vpε(F) :=sup

n

k=

vp(Fk),

where the supremum is taken over all partitionsF=nk=Fkof the setFinto a finite

num-ber of compact pairwiseε-disjoint subsetsFk.

Remark  We writeVp(F) instead ofVp(F). In this case the supremum in the definition

is taken over all partitionsF=nk=Fkof the setFinto a finite number of compact disjoint

subsetsFk.

3.2 Correctness of the definitions

The proofs of measurability of the functions that stay under integral signs (Lemmas -) use ideas from [] (Chapters -) and [] (Chapter ).

Throughout this subsection, we identify each point (r,x) of the space Pd–×Rd–

(r∈Pd–,xRd–) with the linel(r,β), where βd–(r) is a point with coordinates

xwith respect to some orthonormal basis ofd–(r) (the basis ofd–(r) is assumed to continuously change asr∈Pd–changes).

We need the following lemma.

Lemma  For every Borel set B⊂Rd, the setψ(B)of all(r,x)Pd–×Rd– such that

Bl(r,β)=∅is measurable(Pd–×Rd–).The setϕ(B) :={t∈R: B∩Rdt =∅}is measur-able.

Denote byψone of continuous maps from [, ] ontoSd–; such a map exists due to

the Hahn-Mazurkiewicz theorem; see, for example, Theorem . in []. Define a func-tionψ: [, ]×Rd→Pd–×Rd–using the following rule. For each (t,y)∈[, ]×Rd,

letψ(t,y) be the linel(r,β), wherer∈Pd–is the line that containsψ(t), andβ is such

that the linel(r,β) contains the pointy. It is easy to see that the functionψ is

contin-uous. Thenψ(B) =ψ(B), where˜ B˜ := [, ]×B. This means that ψ(B) is a continuous

image of a Borel setB˜ ⊂Rd+ and hence is measurable (see, for example, Theorem 

in []).

(4)

Definition  For a compact setF⊂Rd andε> , denote by N

ε(F) the number ofε -components of the setF;  for an empty set.

Lemma  Let a compact set F⊂Rdbe given.Then,for arbitraryε> ,N

ε(F)is finite and limε→(F) =N(F).

With eachε-componentWof the setF, associate a ball with center in an arbitrary point wW and radius ε

. Balls that correspond to differentε-components ofFare pairwise

disjoint; hence there can be only a finite number of such balls due to the boundedness ofF.

IfFhas a finite number of connected componentsF, . . . ,Fn, thenε:=mini=(Fi,Fj) > 

due to the compactness of components of a compact set. Then(F) =n=N(F) for all

ε<ε.

IfN(F) =∞, then for arbitraryn∈Nwe can choose compact disjoint setsF, . . . ,Fnsuch

thatF=nk=Fk. Then(F)≥nfor allε<mini=(Fi,Fj).

For a compact setF⊂Rdandε> , define functions

vF: Pd–×Rd–→R, vF(r,x) =N

Fl(r,β), () and

F: Pd–×Rd–→R, F(r,x) =Nε

Fl(r,β), NFε:R→R, NFε(t) =

F∩Rdt. The following lemma holds.

Lemma  Let a compact set F⊂Rd andε> be given.The function vε

F is measurable

(Pd–×Rd–).The function Nε

F is measurable.

Consider the (countable) setof all closed balls inRdwith rational centers and radii. Let˜ be the set of all finite unions of the balls from. For eachn∈N, definento be the

family of all sets of the formms=Bis, wheremn, and{Bis}ms=is a collection of pairwise

ε-disjoint sets from˜. Then the setnis countable for eachn∈N.

Note that the functionsF andNFεtake only non-negative integer values. The sets{F≥ }=Pd–×Rd–and{NFε≥}=Rare measurable. Suppose nownis a natural number. Below we prove that

Fn= B=ms=Bisn

m

s=

ψ(BisF)

m s=

ψ(∂BisF)

()

and

NFεn= B=ms=Bisn

m

s=

ϕ(BisF)

m s=

ϕ(∂BisF)

,

(5)

If (r,x) belongs to the right-hand side of (), then there exists a setB=ms=Bisn such thatl(r,β)∩BisFis a non-empty set strictly insideBis,s= , . . . ,m. Since the sets that constituteBareε-disjoint, by the definition ofn, we obtain thatvεF(r,x)mn

and hence (r,x)∈ {

Fn}.

Let (r,x)∈ {Fn}andF, . . . ,Fm,mnbe allε-components of the setl(r,β)∩F. Since

mini=(Fi,Fj) >ε, there existsδ>  such thatmini=(Fi,Fj) >ε+ δ. Consider a finite

coverCof the compact setl(r,β)∩Fby open balls with rational centers and radii such that each ball has diameter less thanδand contains some point froml(r,β)∩F. Denote by Bk the union of closures of all balls from the coverC that intersectFk,k= , . . . ,m.

Thenmk=Bk belongs tonby construction, and hence (r,x) belongs to the right-hand

side of ().

SinceFis a closed set andnis a countable set, Lemma  implies that the sets{vεFn}

and{

Fn}are measurable; hence the functionsvεF andNFεare measurable.

Lemma  The function vF defined by()is measurable(Pd–×Rd–)for each compact

F⊂Rd.

For each fixed (r,x)∈Pd–×Rd–, one haslim

ε→vεF(r,x) =vF(r,x) due to Lemma .

Hence the measurability ofvF is a consequence of the measurability ofvεF,ε> , stated

in Lemma .

Tonelli’s theorem and Lemma  imply that the functionvp(F) is well defined for every

≤p≤ ∞and compactF⊂Rd; hence the functions

p(F) are also well defined for all

ε≥.

3.3 Some properties of the sets variation

The following property is a direct consequence of the definitions.

Property  Letp∈[,∞],ε≥ andF⊂Rdbe a compact set. Thenv

p(F)≤Vpε(F). IfF

isε-connected, thenvp(F) =Vpε(F).

Property  Letp∈[,∞] andF⊂Rdbe a compact set. Then

Vp(F) =lim

ε→V

ε

p(F). ()

From the definition it follows that

p (F)≤Vpε(F) wheneverε>ε. This implies that

limε→Vpε(F) exists and does not exceedVp(F).

Assume thatVp(F) <∞. Then, for arbitraryδ> , there exists a partitionF=

n(δ) k=Fk

of the setFinto pairwise disjoint compact setsFksuch thatVp(F)≤

n(δ)

k=vp(Fk) +δ. Set

ε:=mini=(Fi,Fj). Thenε>  and for allε<εVpε(F)≥Vp(F) –δ. This implies ().

The case when

p(F) =∞can be considered in a similar way.

Property  Letn∈Nand pairwise disjoint compact setsF, . . . ,Fn⊂Rdbe given. Then,

for allp∈[,∞],vp(

n k=Fk)≤

n

k=vp(Fk).

It is sufficient to prove the property in the casen= . SetF:=F∪F.

SinceFandFare compact disjoint sets, we haveρ(F,F) > , and hence, for arbitrary

r∈Pd– andβ d–(r), one hasN(Fl(r,β)) =N(F

(6)

implies thatv(F,r)v(F,r) +v(F,r) for allr∈Pd–, and hencevp(F)≤vp(F) +vp(F) for

allp∈[,∞].

Property  Assumen∈N,ε≥ and pairwiseε-disjoint compact setsF, . . . ,Fn⊂Rdare

given. Then, for allp∈[,∞],

Vpε

n

k=

Fk

=

n

k=

Vpε(Fk).

It is sufficient to prove the property in the casen= . SetF:=F∪F.

Let{Tk}mk=,m∈N, be a partition of the setFinto compact pairwiseε-disjoint subsets.

Then, by Property ,

m

k=

vp(Tk) = m

k=

vp

(TkF)∪(TkF)

m

k=

vp(TkF) + m

k=

vp(TkF)

Vpε(F) +Vpε(F)

since{TkF}mk= and{TkF}mk= are partitions of the setsF andF into pairwiseε

-disjoint compact subsets. This implies that

Vpε(F)≤Vpε(F) +Vpε(F).

On the other hand, for arbitrary partitions of the setsFandFinto compactε-disjoint

sets{Tk}sk=and{Tk}mk=respectively,s,m∈N,

Tksk=Tkmk=

is a partition of the setFinto compactε-disjoint sets, and hence

p(F)≥Vpε(F) +Vpε(F).

Property  If ε≥ andF⊂Rd is a compact set that has exactly nNε-connected

componentsF, . . . ,Fn, then, for allp∈[,∞],Vpε(F) =

n

k=vp(Fk).

Note that eachε-connected component of a compact set is compact. Hence, by Proper-ties  and ,

p(F) =

n

k=Vpε(Fk) =nk=vp(Fk).

Property  IfF⊂Rdis a compact set, α=  andαF:={αx:xF}, then for arbitrary

p∈[,∞]vp(F) =vp(αF) andVp(F) =Vp(αF).

The property follows from the observation that for arbitraryr∈Pd–andβd–(r) one

(7)

4 Variation of a function

4.1 Definition

Definition  Let a setE⊂Rdand a functionf:ERbe given. FortR, the set

L(f;t) :=xE:f(x) =t

is called a level set of the functionf.

The variation of a continuous function is given by the following definition.

Definition  LetE⊂Rd,f:ERbe a continuous on a compact subsetFEfunction,

andp∈[,∞]. Set

vp(f;F) :=

–∞vp

FL(f;t)dt.

IfFis locally connected, then set

Vp(f;F) :=

–∞Vp

FL(f;t)dt.

IfF=E, then we writevp(f) andVp(f) instead ofvp(f;E) andVp(f;E), respectively.

4.2 Correctness of the definitions

We need to prove that the functions under the integral signs are measurable.

Lemma  Let E⊂Rd,f: ERbe a continuous on a compact subset FE function,and

p∈[,∞].Then the function vp(F∩L(f;·))is measurable.

Without loss of generality, we may assume thatRdF=E, and we need to prove that

the functionvp(L(f;·)) is measurable. Consider the graph

(f) :=(x,t)∈Rd+:xE,f(x) =t ()

of the functionf and two functionsv(f): Pd×Rd→Randvf:R×Pd–×Rd–→R

defined by the formulavf(t,r,x) =vL(f;t)(r,x) (see () for the definition of the functionvF).

Since the setEis compact and the functionf is continuous onE, the set(f)⊂Rd+is

compact. This, by Lemma , implies that the functionv(f)is measurable (Pd×Rd).

Recall thatRd+={(x, )∈Rd+:x∈Rd}. The function

φ:R×Pd–×Rd–→Pd×Rd∩(R,y)∈Pd×Rd: R⊂Rd+

that maps a point (t,r,x) to the point (R,y) withR={(z, ),zr}andy= (x,t) is continuous and has continuous inverse. Moreover, for arbitrary (t,r,x)∈R×Pd–×Rd–, we obtain

vf(t,r,x) =v(f)(φ(t,r,x)). Hence, for arbitraryc∈R, the functionφmaps the set

(t,r,x)∈R×Pd–×Rd–: vf(t,r,x)c

(8)

and the set

(R,y)∈Pd×Rd:v(f)(R,y)c

∩(R,y)∈Pd×Rd:R⊂Rd+.

The latter is an intersection of a measurable (due to measurability ofv(f)) set and a closed

set; hence the former is also a measurable set. This means that the functionvf is

mea-surable (R×Pd–×Rd–), and hence the statement of the lemma is true due to Tonelli’s

theorem.

The following result is well known (see, for example, [], Lemma , or [], Lemma  in Chapter ).

Lemma  For an arbitrary function f: E→R,denote by Textrthe set of t∈Rsuch that

L(f;t)contains an extremum point.Then Textris at most countable.

Lemma  Let E⊂Rd,f:ERbe a continuous on a compact locally connected subset

FE function,and p∈[,∞].Then the function Vp(F∩L(f;·))is measurable.

Without loss of generalization, we may assume thatF=E. Taking into account Prop-erty , it is sufficient to prove that each of the functions

p(L(f;·)) is measurable,ν> .

Let ν >  be fixed. The function Nf(t) := (L(f;t)) =Nν(f)(t) is measurable due to Lemma  ((f) is defined by ()). Due to Lemma , for eachk= , , . . . , the setTk:=

{Nf=k} \Textris measurable; moreover, obviously these sets are pairwise disjoint.

Letc∈Rbe given. Then

VpνL(f;·)≤c\Textr=

k∈N

tTk:Vpν

L(f;t)c

and it is sufficient to prove that for eachk∈Nthe set{tTk:Vpν(L(f;t))c}is

measur-able.

Let somek∈Nandt∗∈Tkbe fixed. The setL(f;t∗) contains exactlyk ν-connected

componentsF, . . . ,Fk. Each of the components is a compact set, hence there existsε> 

such that the setsUs(ε) ={x:ρ(x,Fs) <ε},s= , . . . ,k, are pairwiseν-disjoint. SetU(ε) :=

k s=Us(ε).

There existsδ>  such thatL(f;t)U(ε) for allt∈(t∗–δ,t∗+δ). Really, assume the contrary, suppose that there exists a sequencean,an→ asn→ ∞, such that each of the

level setsL(f;t∗+an) contains a pointxn∈/U(ε). Switching to a subsequence, if needed,

we may assume that the sequencexnconverges to somex∈Rd. SinceU(ε) is an open set,

x∈/U(ε). However, this is impossible sincef is continuous, and hencef(x) =t∗.

For eachs= , . . . ,k, consider an arbitrary pointxsFs. Since the level setL(f,t∗) does

not contain extremums,Fis locally connected andf is continuous; for small enoughεs>

, the setf(Bd(x

s,εs)) contains a neighborhood (t∗–δs,t∗+δs) oft∗ (δs> ). Hence, for

arbitraryt∈(t∗–δs,t∗+δs), the level setL(f;t) contains at least oneν-component inside

Us(ε). This means that there existsδ=δ(t∗) >  such that each of the level setsL(f,t),

t∈(t∗–δ,t∗+δ) contains at least oneν-component insideUs(ε),s= , . . . ,k. Hence, for all

(9)

s= , . . . ,k. This implies that for alltTk∩(t∗–δ,t∗+δ)

VpνL(f;t)=

k

s=

vp

L(f;t)Us(ε)

()

due to Property .

For eacht∗∈Tk, setW(t∗) := (t∗–δ(t∗),t∗+δ(t∗)); the setsW(t∗),t∗∈Tk, constitute

an open cover of the setTk. SinceRis a Lindelöf space, we can find a countable subcover

W,W, . . . ofTk. SetW˜:=W∩Tk,W˜m:= (Wm\

m–

s= Ws)∩Tk,m= , , . . . . We obtain

a countable partition of the setTkinto pairwise disjoint measurable subsetsW˜m,m∈N,

such that on each of the setsW˜m we have representation () of the functionVpν(L(f;·))

(the setsUs(ε),s= , . . . ,k, might be different for differentm∈N). Due to Lemma , this

implies the measurability of the set

tTk:Vpν

L(f;t)c=

m∈N

˜ Wm

k

s=

vp

L(f;t)Us(ε)

c

and hence the lemma is proved.

4.3 Some properties of the function variation

Below we list some properties of the function variation. Everywherep∈[,∞] and a com-pact setF⊂Rdare fixed,f is continuous onFfunction. For properties ofV

p(f), the setF

is further assumed to be locally connected.

Property  vp(f) andVp(f) are non-negative. Iff is constant, thenvp(f) =Vp(f) = .

The fact that variations are non-negative follows from the definition. Iff is constant, then it has exactly one non-empty level set.

Property  Ifα,β∈R, thenvp(α·f+β) =|α|vp(f) andVp(α·f +β) =|α|Vp(f).

Note thatL(α·f+β;t) =L(f;α–·(t–β)) for allα=  andβ,t∈R. Making substitution s=α–·(t–β) in the integrals from Definition , we obtain the required equalities. In the caseα= , the property follows from Property .

Property  For arbitraryα= ,vp(f;F) =vp(f(α·);α–F) and

Vp(f;F) =Vp

f(α·);α–F.

The property follows from Property .

Property  Lett∈Rbe such thatF\L(f;t) has exactlyn≥ connected components F,F, . . . ,Fn. Assume that for allk= , . . . ,n,Fk\FkL(f;t). Thenvp(f;F)

n

k=vp(f;Fk)

andVp(f;F) =

n

k=Vp(f;Fk).

Consider arbitrarys∈R,s=t. Fork= , . . . ,n, setWk:=FkL(f;s). Then the setWkis

closed,k= , . . . ,n, andWkFk due to conditions of the property and the fact that

(10)

disjoint sets. From Properties  and  it follows thatVp(L(f;s)) =

n

k=Vp(FkL(f;s)) and

vp(L(f;s))

n

k=vp(FkL(f;s)). The statement of the property now follows from

Defini-tion .

Property  Ifd=  andf: [, ]→Ris a continuous function, then for allp∈[,∞]

vp

f; [, ]=Vp

f; [, ]=

f.

Remark fis the classical variation of a univariate functionfon [, ]. We allowf to be +∞in the case whenf is not a function of bounded variation.

The Banach indicatrix theorem [] states thatf is equal to the integral overt∈R of number of points inL(f;t). In the definition ofvp(f; [, ]), the number of components

ofL(f;t) is integrated overt∈R. Each component of a compact set inRis a point or a segment. The family of level setsL(f;t) that contain a segment as a connected component is at most countable because each of such level sets contains extremum (see Lemma ). Hencevp(f; [, ]) =

f. It is easy to see thatvp(f; [, ]) =Vp(f; [, ]).

Property  Letϕ: [, ]→Rbe a continuous function and d∈N. Let: Bd→R, (x) =ϕ(|x|). Then, for allp∈[,∞],

vp

;Bd

=Vp

;Bd

= ·

ϕ.

In the cased= , the property follows from Property , so we can assume thatd≥. Let arbitraryt=ϕ() be fixed. For arbitraryr∈Pd–andβd–(r) the numberN(L(fϕ;t)l(r,β)) can be obtained by the following procedure: consider the liner=l(r,θ) and mark points of the setL(fϕ;t)l(r,θ); cut the interval (–|β|,|β|) from the line and stick the points –|β|and|β|together; the number of components of marked points on the obtained ‘cut’ line is equal toN(L(fϕ;t)l(r,β)). This shows that for arbitraryβ

NL(fϕ;t)l(r,β)

NL(fϕ;t)l(r,θ)

. ()

From the choice oftit follows that

θ∈/L(fϕ;t), ()

and hence there exists ε>  such that B(ε)∩L(fϕ;t) =∅. This implies that the set L(fϕ;t)l(r,θ) does not contain pointsxwith|x|<εand hence for allβsuch that|β|<ε () becomes equality. This implies thatv(L(fϕ;t),r) =N(L(fϕ;t)l(r,θ)). From () it follows thatN(L(fϕ;t)l(r,θ)) = ·N(L(ϕ;t)). Equalityvp(fϕ;B) = ·

ϕfollows from Property 

now. Equalityvp(fϕ;B) =Vp(fϕ;B) follows from the geometry of the level sets offϕ.

5 Ostrowski type inequalities

5.1 Auxiliary results

Lemma  Let d∈N,d≥,ε> ,x∈Rd,rPd–and a measurable set FBd(x,ε)be

(11)

and r such that

μd–βd–(r) :Fl(r,β)=∅>A·μd–Bd–(ε) wheneverμdF>α·μdBd(ε).

The fact thatαdoes not depend onεfollows from the observation that

μdF

μdBd(ε)=

μd(

εF)

μdBd

and

μd–{βd–(r) :Fl(r,β)=∅}

μd–Bd–(ε) =

μd–{βd–(r) :

εFl(r,β)=∅}

μd–Bd– .

The fact thatαis independent ofxandris obvious. The existence ofαfollows from the equality

μdF=

d–(r)Bd(y,ε)

μl(r,β)∩Fμd–(dβ), ()

where yd–(r) is such that the linel(r,y) containsx and equality μd–(d–(r)∩ Bd(y,ε)) =μd–Bd–.

Lemma  Let p∈[,∞),A> and B∈[,A]be given.Then

A

B+ p(A–B)

 – B A

p

.

It is sufficient to prove that the functionϕ(x) = p+ ( – p)x– ( –x)pis non-negative

on [, ]. Sinceϕ() =ϕ() = , the functionϕhas at least one zero on (, ). The function

ϕ(x) =p( –x)p–+  – pis decreasing on [, ], hence has at most one zero on (, ). This

implies thatϕ() >  and hence the functionϕis increasing on [,x∗] and is decreasing on [x∗, ], wherex∗is zero ofϕon (, ); henceϕis non-negative on [, ].

5.2 Main results

The following theorem is the main tool to prove Ostrowski type inequalities for functions and sets of bounded variation below.

Theorem  Let d∈Nand two sets F,WBdbe given.Assume that the following

proper-ties hold:

. Fis measurable andθ∈/F;

. Wis closed andθ∈/W;and

. IfxFandyBd\F,thenxyW=.

Then,for all p∈[,∞],

μdFμ

dBd

(12)

The inequality is sharp in the sense that for arbitraryε> there exist sets F and W that satisfy conditions above and such that

μdF>

μdBd  –ε

vp(W).

If ()becomes equality,thenμdF= .

We will prove Theorem  in the next subsections. Here we state two consequences of this theorem, which can be considered as Ostrowski type inequalities.

Theorem  Let d∈Nand a continuous function f: Bd→Rbe given.Then,for all p∈ [,∞],

μdBd

Bd

f(x)dx–f(θ)≤vp(f)  .

The inequality is sharp.It becomes equality only in the case when f is constant.

Due to Property , we can assume thatf(θ) = , and it is sufficient to prove that

Bdf(x)dx

μdBd

vp(f). ()

Consider a set

:=(x,t)Bd×[,∞) :f(x)≥t.

Then

Bd

f(x)dxμd+=

t≥

μd∩Rdt+dt. ()

For eacht> , consider the setsF:=∩Rd+

t andW:=(f)∩Rdt+(see () for the

defi-nition of(f)). BothFandWare closed sets that do not containθsincef(θ) = . IfxF andyBd\F, thenf(x)≥tandf(y) <tand hence the segmentxycontains a pointzwith f(z) =t, i.e.,xyW=∅. This means that all the conditions of Theorem  are satisfied and hence

μd∩Rdt+=μd(F)≤μ

dBd

vp(W) =

μdBd

vp

L(f;t)

with equality possible only in the case whenμdF= . Taking into account (), we obtain

μd+μ

dBd

t≥

vp

L(f;t)dtμ

dBd

t∈Rvp

L(f;t)dt=μ

dBd

vp(f)

(13)

For allε> , consider the functionϕε: [, ]→R,ϕε(t) =  fortε,ϕε() =  andϕεis linear on [,ε]. Due to Property , for the radial function(x) :Bd→R,(x) =ϕε(|x|), and arbitraryp∈[,∞]vp(fε) = ; moreover,

Bdfε(x)dxμdBdasε→. This proves the sharpness of the stated inequality.

Theorem  Let d∈Nand a closed set FBdbe given.Ifθ/F,then for all p[,]

μdFμ

dBd

vp(F).

The inequality is sharp.If equality holds,thenμdF= .

It is enough to apply Theorem  with W =F; all three conditions of Theorem  are satisfied.

For arbitraryε> , consider a set:={xBd: |x| ≥ε}. For allp∈[,∞],vp(Fε) = ;

μdF

εμdBdasε→. This proves that the stated inequality is sharp.

Remark  In all three theorems variationvpcan be substituted byVpdue to Property .

The inequalities will remain sharp.

Remark  Properties  and  state thatVpis additive. This gives motivation to callVpa

variation, rather thanvp.

5.3 More auxiliary results

Denote byF˜the set of all pointsxFsuch thatlimδ→+μ

d(FBd(x,δ))

μdBd(δ) = . ThenF˜∩Sd–=∅ and, by the Lebesgue density theorem,

μdF˜=μdF. ()

Lemma  Assume that the conditions of Theoremhold.If r∈Pd–is such that v(W,r) = ,then for arbitraryβd–eitherF˜intBdl(r,β),orF˜l(r,β) =.

Assume that for someβd–(r) there existx∈ ˜Fl(r,β) andy∈(intBdl(r,β))\ ˜F. From the definition ofF˜it follows that there exista>  and a sequenceρn→ asn→ ∞

such thatμd(Bd(y,ρ

n)\F)a·μdBd(ρn) for alln∈N. From () (withF substituted by

Bd(y,ρ

n)\F) it follows that there existsA>  such that

μd–(ρn) >A·μd–Bd–(ρn) ()

for alln∈N, where

(ρn) =

βd–(r) :Bd(y,ρn)\F

l(r,β)=∅.

Sincex∈ ˜F, there existsδ>  such that for allρ<δone hasμd(Bd(x,ρ)∩F)α( –A)·

μdBd(ρ) (the numberα( –A) is defined in Lemma ). Lemma  implies that

(14)

for allρδ, where

(ρ) =

βd–(r) :Bd(x,ρ)∩Fl(r,β)=∅.

Choosenso big thatρn<δ. Then

μd–(ρn) +μd–(ρn) >μd–Bd–(ρn) ()

due to () and (). Moreover, sincex,yl(r,β), we receive that

(ρn),(ρn)⊂d–(r)∩Bd(β,ρn) ()

and

μd–d–(r)∩Bd(β,ρn)

=μd–Bd–(ρn)

. ()

Set=(ρn)∩(ρn). Then, due to (), () and (),μd–> . But each linel(r,β),

β, contains a point fromW due to Condition  of Theorem  and the definitions of the sets(ρn) and(ρn); this contradicts assumptionv(W,r) =  of the lemma.

Lemma  Assume that the conditions of Theoremhold. Let R⊂ Pd– be such that

v(W,r) = for all rR.If R contains d lines that are not contained in any d– -dimensional hyperplane,thenμd(F) = .

Due to () it is enough to prove thatF˜ =∅. Letr, . . . ,rdbe the lines from the

state-ment of the lemma, and let ρ, . . . ,ρd be unit vectors parallel to these lines. Set P :=

{d

k=tkρk:tk∈(–, ),k= , . . . ,d}, thenPis an open inRdset.

Consider arbitraryx∈intBd. Chooseε>  such thatx+εPBd. Then, for all points yfrom the segmentθx,Py:=y+εPBd.yθxPyis an open cover of a compact setθx,

hence it contains a finite subcoverP,P, . . . ,Pm,m∈N. From Lemma  it follows that for

eachs= , . . . ,meitherPs⊂ ˜F, or

Ps∩ ˜F=∅. ()

Sinceθ∈ ˜/F, we obtain that () holds for eachs= , . . . ,mand hencex∈ ˜/F.

5.4 Proof of Theorem 1

Ifv(W,r)≥ for almost allr∈Pd–, thenv

p(W)≥ and inequality () holds. It is strict

because Condition  of Theorem  holds. If there is a setR⊂Pd–of positive measure such

thatv(W,r) =  for allrR, thenμdF=  due to Lemma , and inequality () holds.

Assume that there exists R⊂Pd–, μR> , such that v(W,r) =  for all rR and v(W,r)≥ for almost allr∈Pd–\R. Then

vp(W)≥ –

μR

(15)

Really, ifp=∞, thenv∞(W)≥ in the caseμR<μPd–andv∞(W) =  in the caseμR=

μPd–=μSd–. In both cases () holds. In the casep[,)

vp(W)≥

μSd–

μR+ p·μSd––μRp

≥ – μR

μSd–

due to Lemma .

Conditions  and  of the theorem imply that there existsε>  such thatBd(ε)W=

andBd(ε)F=. Set:=

rR(r∩Bd). Below we prove that

μd(F) <μ

d

 . ()

In order to prove (), it is enough to show that

μd(∩ ˜F) <μ

d

 ()

due to (). Consider arbitraryrR. Then all points of the intersectionr∩ ˜Fare from one side ofrBd(ε). This fact can be proved using arguments similar to the proof of Lemma .

Denote byχthe characteristic function of the set∩ ˜F. Thenχ(x) =  for all|x|<εand

χ(x) +χ(–x)≤ for allx. This implies (). Finally, having (), we can write

μdFμd(F∩) +μdBd\

<μdBd– μ

d

=μdBd– ·

μdBd

μSd–μR

= μ

dBd

 – μR

μSd–

.

The latter together with () proves ().

The same example as in Theorem  shows that inequality () is sharp.

Acknowledgements

The author was partially supported by grant 0117U001208. Competing interests

The author declares that they have no competing interests. Author’s contributions

All authors made an equal contribution to the paper, have read and approved the final manuscript.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Received: 16 March 2017 Accepted: 16 June 2017

References

(16)

2. Dragomir, SS: A companion of Ostrowski’s inequality for functions of bounded variation and applications. Int. J. Nonlinear Anal. Appl.5(1), 89-97 (2014)

3. Clarkson, JA, Adams, CR: On definitions of bounded variation for functions of two variables. Trans. Am. Math. Soc.

35(4), 824-854 (1933)

4. Giorgi, E: Su una teoria generale della misuran– 1-dimensionale in uno spazio ardimensioni. Ann. Mat. Pura Appl.

36(4), 191-213 (1954)

5. Vitushkin, AG: On Multidimensional Variations. Gosudarstv. Izdat. Tehn.-Teor. Lit., Moscow (1955) (in Russian) 6. Ivanov, LD: Variations of the Sets and Functions. Nauka, Moskow (1975) (in Russian)

7. Sagan, H: Space-Filling Curves. Springer, New-York (1994)

References

Related documents