REPRESENTATIONS OF MULTIMEASURES VIA THE

MULTIVALUED BARTLE-DUNFORD-SCHWARTZ
INTEGRAL
arXiv:2102.09886v3 [math.FA] 13 Feb 2023

LUISA DI PIAZZA, KAZIMIERZ MUSIAL, ANNA RITA SAMBUCINI

Abstract. An integral for a scalar function with respect to a multi-

measure N taking its values in a locally convex space is introduced. The
definition is independent of the selections of N and is related to a func-
tional version of the Bartle-Dunford-Schwartz integral with respect to a
vector measure presented by Lewis. Its properties are studied together
with its application to Radon-Nikodým theorems in order to represent
as an integrable derivative the ratio of two general multimeasures or two
dH -multimeasures; equivalent conditions are provided in both cases.

1. Introduction
The theory of the Bartle-Dunford-Schwartz-integral (BDS-integral) of a
scalar function with respect to a vector valued measure was introduced in
1955 by R. C. Bartle, N. Dunford and J. T. Schwartz [1], and subsequently
extensively studied by several authors ( [14, 19–22, 27, 29, 30]).
In ‘70 Lewis [19] proposed an equivalent functional version of the integral
(see also [20,21,27]). The literature concerning the Bartle-Dunford-Schwartz
integration is rather wide so we quote here only those papers which are close
to the topic of our work and two books ( [21, 27]).
In [22] the second author proved the existence of the Radon-Nikodým deriv-
ative of a vector valued measure ν with respect to a vector valued measure
κ by means of the BDS-integral, under suitable hypotheses on the measures
ν and κ. From this article then came out [4], where the derivative belongs
to a suitable space and that of [12], where a Radon-Nikodým theorem was
given bases on a construction of Maynard type.
In [18] Kandilakis defined an integral of a scalar function with respect
to a multimeasure N , whose values were weakly compact and convex sub-
sets of a Banach space. This integral is constructed using the selections of
N . Contrary to the classical BDS-integral, the Kandilakis’ integral is only
sublinear with respect to the integrable functions.
It is our aim to define an integral with respect to a multimeasure N ,
independent of the selections of N . To achieve it we take into account
2020 Mathematics Subject Classification. Primary 28B20; Secondary 26E25, 26A39,
28B05, 46G10, 54C60, 54C65.
Key words and phrases. Locally convex space, multifunction, Bartle-Dunford-Schwartz
integral, support function, selection, Radon-Nikodým theorem.
1
2 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

the support functions of N . Our approach is a consequence of Kandilakis’

calculations. The considered multimeasures N are very general, in fact they
can take nonempty closed, convex values in an arbitrary locally convex space
X. Sometimes we will assume quasi-completeness of X (closed bounded sets
are complete).
The theory of multifunctions and multimeasures is an interesting field
of research since it has applications in various applied sciences. In partic-
ular, recently, interval-valued multifunctions and multimeasures have been
applied also to signal and image processing, see for example [10] and the
references therein.
In Section 2 we study properties of the integral. Moreover, we compare
the Aumann-BDS integral and the new one: if the multimeasure N possesses
selections and the scalar integrable function is bounded, then the Aumann-
BDS integrability implies the N -integrability. To obtain the main results we
require the existence of control measures for the multimeasures under inves-
tigation and in Section 3 we give conditions guaranteeing the existence of
such controls. In particular in Theorem 3.5 a characterization of its existence
through the countable chain condition (ccc) is provided, while in Theorem
3.1 we describe the class of Banach spaces with all cb(X)-multimeasures
being dH -multimeasures (and so admitting control measures). In Theorem
3.6 a class of spaces is given in which every c(X)-valued multimeasures
has (ccc). We quote also very recent characterization of Rodriguez [28] of
Banach spaces with all multimeasures admitting control measures. As an
application of the new integral, we study the problem of existence of the
Radon-Nikodým derivative of a given multimeasure M with respect to a
multimeasure N . In the classical measure theory, the Radon-Nikodým theo-
rem states in concise conditions, namely absolute continuity, domination
and subordination, how a measure can be factorized by another measure
through a density function. In Section 4 we first consider arbitrary multi-
measures. In such a case we are able to characterize the existence of the
Radon-Nikodým derivative of M with respect to N (see Theorem 4.4) by
means of the notions of uniform scalar absolutely continuity, uniform scalar
domination and uniform scalar subordination. In Section 5 multimeasures
which are countably additive in the Hausdorff metric are studied. In such
a case the multimeasures take values in nonempty, bounded, closed, convex
subsets of a locally convex space. The differentiation of M with respect to
N is in general not equivalent to differentiation of its Rådström embedding
j ◦ M with respect to j ◦ N . The reason is that one has to take into ac-
count also integration with respect to the measure j ◦ (−N ). In the case
of multimeasures, we have in general j ◦ (−N ) 6= −j ◦ N ; moreover, since
the proofs given are based on support functions which obviously lost the
additivity, the methods of our proofs are not a mere repetition of the vector
case. The main result is Theorem 5.6 where we find conditions (the strong
uniform scalar absolutely continuity, the strong uniform scalar domination
and the strong uniform scalar subordination) guaranteing the existence of
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 3

the Radon-Nikodým derivative both of M with respect to N and of their

Rådström embeddings. At last in Section 6 we provide some examples of
multimeasures that can be represented by another multimeasure through a
multivalued Bartle-Dunford-Schwartz density.

2. Preliminaries
Throughout (Ω, Σ) is a measurable space, the real numbers are denoted
by R and R+ 0 denotes the non-negative reals. If ν : Σ → (−∞, +∞] is a
measure, then |ν| denotes its variation. ΣE is the family of all Σ-measurable
subsets of E. Let X be a locally convex linear topological space (shortly,
locally convex space) and let X ′ be its conjugate space. Given a subset S of
X, we write co(S), aco(S) and span(S) to denote, respectively, the convex,
absolutely convex and linear hull of S.
The symbol c(X) denotes the collection of all nonempty closed convex sub-
sets of X and cb(X), cwk(X), ck(X) denote respectively the family of all
bounded and the family of all (weakly) compact members of c(X). For every
C ∈ c(X) the support function of C is denoted by s(·, C) and defined on
X ′ by s(x′ , C) = sup{hx′ , xi : x ∈ C}, for each x′ ∈ X ′ . The symbol ⊕
denotes the closure of the Minkowski addition.
We say that M : Σ → c(X) is a multimeasure if for every x′ ∈ X ′ the
set function s(x′ , M (·)) : Σ → (−∞, +∞] is a σ-finite measure. The σ-
finiteness of each s(x′ , M ) seems to be the weakest possible assumption,
otherwise we meet problems connected with the RN-theorem. Simply if
some ν := s(x′ , M ) is not σ-finite but is absolutely continuous with respect
to a finite measure µ, then there is a set Ω0 such R that ν is σ-finite on Ω0
and takes only values 0, +∞ on Ω c . If ν(E) = E f dµ, then f must take
infinite values and we are not interested in such a situation.
Given a multimeasure M : Σ → c(X) we denote by N (M ) := {E ∈
Σ : M (E) = {0}} the family of null sets. A multimeasure M : Σ → c(X) is
said Sto be σ-bounded if there is a sequence (Ωn )n of elements of Σ such that
Ω \ n Ωn ∈ N (M ) and M is cb(X)-valued on each algebra ΣΩn .
A multimeasure M is called pointless if its restriction to no set E ∈
Σ \N (M ) is a vector measure. Each multimeasure determined by a function
(see [7] for the definition) that is not scalarly equivalent to zero function is
pointless. Less trivial examples can be deduced from [24, Example 1.11] if
one assumes that the function r appearing there is strictly positive. Two
multimeasures M, N : Σ → c(X) are consistent if there exists H ∈ Σ such
that M and N are pointless on H and vector measures on H c .
If M is a cb(X)-valued multimeasure, then for each x′ ∈ X ′ the measure
s(x′ , M ) is finite. If A ∈ Σ, then M |A is the multimeasure defined on ΣA by
(M |A )(E) := M (A ∩ E). A multimeasure M : Σ → c(X) is called positive,
if 0 ∈ M (E) for each E ∈ Σ.
4 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

If X is a Banach space and a multimeasure M : Σ → cb(X) is countably

additive in the Hausdorff metric dH , then it is called a dH -multimeasure.
In the following if Z is any metric space, we use the symbol BZ to denote
its closed unit ball.
A helpful tool to study the dH -multimeasures is the Rådström embedding j :
cb(X) → ℓ∞ (BX ′ ), defined by j(A) := s(·, A), (see, for example [2, Theorem
3.2.9 and Theorem 3.2.4(1)] or [9, Theorem II-19]) It is known that BX ′
can be embedded into ℓ′∞ (BX ′ ) by the mapping x′ −→ ex′ , where hex′ , hi =
h(x′ ), for each h ∈ ℓ∞ (BX ′ ). Moreover, the range of BX ′ is a norming subset
of ℓ′∞ (BX ′ ). The embedding j satisfies the following properties:
2.a): j(αA ⊕ βC) = αj(A) + βj(C) for every A, C ∈ cb(X), α, β ∈ R+ ;
2.b): dH (A, C) = kj(A) − j(C)k∞ , A, C ∈ cb(X);
2.c): j(cb(X)) is a closed cone in the space ℓ∞ (BX ′ ) equipped with the
norm of the uniform convergence.
Observe that instead of ℓ∞ (BX ′ ), we may use CB (BX ′ , τ(X ′ ,X) ) (where τ(X ′ ,X)
is the Mackey topology) for weakly compact sets, C(BX ′ , σ(X ′ , X)) in case
of compact sets and CB (BX ′ , k · k∞ ) in case of closed bounded sets. But as
we do not apply any special properties of space in which we embed cb(X),
we will stay with

(1) ℓ∞ (BX ′ ).

We denote by R(ME ) the range of the multimeasure M in c(X), restricted to

measurable subsets of E: R(ME ) := {M (F ) : F ∈ ΣE } ⊂ c(X). Moreover,
R(ME ) := {z ∈ X : ∃ F ∈ ΣE , z ∈ M (F )} ⊂ X.
We say that a multimeasure M : Σ → c(X) is absolutely continuous with
respect to a multimeasure N : Σ → c(X) (we write then M ≪ N ), if
N (N ) ⊂ N (M ). If M ≪ N and N ≪ M , then the multimeasures are called
equivalent.
We say that a non negative measure µ is a control measure for a mul-
timeasure M : Σ → c(X) if µ is a finite measure and for each E ∈ Σ
the condition µ(E) = 0 yields M (E) = {0}. If µ is a control measure for
M : Σ → c(X), then M is σ-bounded if and only if it is locally bounded, i.e.
for each set E ∈
/ N (µ) there exists a subset F of E of positive µ-measure such
that M is cb(X)-valued on ΣF . It is a direct consequence of [1, Theorem
1.4] that each dH -multimeasure has a control measure.
If M : Σ → c(X) is a multimeasure, then SM denotes the family of all
countably additive X-valued selections of M .

Definition 2.1.
• Let n : Σ → X be a vector measure. A measurable function f :
Ω → R is called Bartle-Dunford-Schwartz (BDS) integrable with
respect to n, if for every x′ ∈ X ′ f is hx′ , ni-integrable and for
each E ∈ Σ there exists a point ν(E) ∈ X such that hx′ , ν(E)i =
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 5
Z
f dhx′ , ni, for every x′ ∈ X ′ . Such an approach to the Bartle-
E
Dunford-Schwartz integral was suggested by Lewis [19].
• Let N : Σ → c(X) be a multimeasure. A measurable function
f : Ω → R is called Aumann-Bartle-Dunford-Schwartz (Aumann-
BDS)-integrable with respect to N if SN 6= ∅ and f is integrable in
the sense of Bartle-Dunford-Schwartz with respect to all members
of SN . The integral on a set E ∈ Σ is then defined by the formula
Z  Z 
(s) f dN := (BDS) f dn : n ∈ SN .
E E

The above definition was suggested by Kandilakis [18] for cwk(X)-

valued multimeasures and a Banach space. In that case the set in
the braces was closed and the additional closure is superfluous. The
symbol (s) used here indicates that the integral is constructed using
the selections of N .
Remark 2.2.
• A multimeasure N : Σ → c(X) is called rich if N (A) = {n(A), n ∈ SN }.
By a result of Costé, quoted in [16, Theorem 7.9], this is verified for
multimeasures which take as their values weakly compact convex
subsets of a Banach space X, or cb(X)-valued when X is a Banach
space possessing the RNP ( [11, Théorème 1]).
• If N is a cwk(X) valued multimeasure of bounded variation and X
is a Banach space, then SNZ 6= ∅ (see for example [15,17]). Moreover,
by [18, Theorem 3.2], (s) f dN ∈ cwk(X).
E
• If f, g are Aumann-BDS-integrable, then
Z Z Z
(s) (f + g) dN ⊆ (s) f dN ⊕ (s) g dN .
E E E
The example of g = −f shows that the equality fails in general. In
particular, the integral in not additive.
• We know from [18, Theorem 3.3] that in case of a multimeasure N
with cwk(X) values in a Banach space X and bounded f , we have
for every x′ ∈ X ′ and every E ∈ Σ
 Z  Z Z
′ + ′
(2) s x , (s) f dN = f ds(x , N ) + f − ds(−x′ , N )
E
ZE ZE
+ ′
= f ds(x , N ) + f − ds(x′ , −N )
E E
in general
Z Z
6= f + ds(x′ , N ) − f − ds(x′ , N )
ZE E
′
= f ds(x , N ) .
E
6 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

But it follows from that proof that the equality (2) holds true in
each case when SN 6= ∅, that is even when X is a locally convex
space and N is not weakly compactly valued.
In particular, if f ≥ 0, then
 Z  Z
′
(3) s x , (s) f dN = f ds(x′ , N )
E E
and the integral is additive for non-negative functions. Moreover,
it follows from (2) that the integral is a multimeasure, for every
Aumann-BDS-integrable f .
As noticed in [30, Remark 3.2], if f is negative, then the equality (3) may
fail. In fact, according to (2), we have then
 Z  Z
′
s x , (s) (−1) dN = 1 ds(−x′ , N ) = s(−x′ , N (E))
E E
in general
Z
′
6= −s(x , N (E)) = −1 ds(x′ , N ) .
E
Now we would like to define an integral with respect to a multimeasure that
would be independent of selections of the multimeasure but would be consis-
tent with earlier definitions via selections. We know already ( [18,30]) that if
X is a Banach space and a non-negative θ : Ω → R is integrable with respect
to N : Σ → cwk(X), then for every E ∈ Σ there exists WEZ ∈ cwk(X) such
that for every x′ ∈ X ′ holds true the equality s(x′ , WE ) = θ ds(x′ , N ).
E

We take this property as the definition of the integral of a non-negative

function with respect to an arbitrary multimeasure N : Σ → c(X). Our
approach is close to Lewis’ [19] functional equivalent definition of an integral
with respect to a vector measure. It is worth to remember that the Lewis
definition in [19] was also considered by Kluvánek in [20] (see also [27]). We
call the integral that we are going to define a multivalued-Bartle-Dunford-
Schwartz integral since Bartle, Dunford and Schwartz were the first who
considered such a kind of integration in the vector case.
Definition 2.3. Let N : Σ → c(X) be a multimeasure. If f : Ω → R
is a non-negative measurable function, we say that f is multivalued-Bartle-
Dunford-Schwartz integrable with respect to N (shortly BDSm - integrable
with respect to N ) in c(X) (cb(X), cwk(X), ck(X)), if for every E ∈ Σ there
exists CE ∈ c(X), (cb(X), cwk(X), ck(X)) such that for every x′ ∈ X ′
Z
(4) s(x′ , CE ) = f ds(x′ , N ).
E
Z
We set CE := f dN .
E
We say that a measurable f : Ω → R is multivalued–Bartle-Dunford-Schwartz
integrable with respect to N in c(X) (cb(X), cwk(X), ck(X)), if f + and f −
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 7

are BDSm -integrable with respect to N in c(X) (cb(X), cwk(X), ck(X)).

Then, for every E ∈ Σ, we define the integral of f with respect to N as
Z Z Z
f dN := f + dN ⊕ f − d(−N ) .
E E E
The above definition is consistent with the property described in [18, Theo-
rem 3.3] (There is a missprint in [18, Theorem 3.3], the sign minus should
be replaced by plus). Equivalently, f : Ω → R is BDSm -integrable with re-
spect to N in c(X), (cb(X), cwk(X), ck(X)), if for each E ∈ Σ there exists
Mf (E) ∈ c(X) (cb(X), cwk(X), ck(X)) such that for every x′ ∈ X ′
Z Z
′ + ′
(5) s(x , Mf (E)) = f ds(x , N ) + f − ds(x′ , −N )
E E
Z
and the right hand side of (5) makes sense. We write then f dN :=
E
Mf (E).
The above definition is consistent with the property described in [18,
Theorem 3.3]. Since s(x′ , ±N ) : Σ → (−∞, +∞] are σ-finite measures for
every x′ then, by (5), the set functions s(x′ , Mf ) are σ-finite measures with
values in (−∞, +∞]. Consequently, Mf is a multimeasure. Let us notice
that if N is a vector measure (say N = ν) then the right hand side of (5)
looks as follows: Z
′
s(x , Mf (E)) = f dhx′ , νi .
E
One can easily check that the integral on the right hand side has to be finite
and M is a vector measure (see Lemma 2.7).
Remark 2.4.
• Assume that SN 6= ∅ (see for example Remark 2.2). Observe that
if f is a bounded, measurable, Aumann-BDS-integrable function
whose integral belongs to c(X) then f is BDS m -integrable with
respect to N and
Z
(s) f dN = Mf (E).
E
• If f is a BDS m -integrable with respect to N function then Mf (A ∪
B) = Mf (A) ⊕ Mf (B) for every A, B ∈ Σ with A ∩ B = ∅. In fact,
for every x′ ∈ X ′ , we have:
 Z  Z Z
′
s x, f dN = (f χA∪B ) ds(x , N ) + (f χA∪B )− ds(x′ , −N )
+ ′
A∪B
ZΩ Ω
Z
= (f χA + f χB ) ds(x , N ) + (f − χA + f − χB ) ds(x′ , −N )
+ + ′
Ω Ω
 Z   Z 
= s x′ , f dN + s x′ , f dN = s(x′ , Mf (A) ⊕ Mf (B)).
A B
8 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

So, for every x′ ∈ X ′ , it is s(x′ , Mf (A ∪ B)) = s(x′ , Mf (A)) +

s(x′ , Mf (B)) = s(x′ , Mf (A) ⊕ Mf (B)).
Remark 2.5. Let X be a Banach space and N : Σ → cb(X) be a dH -
multimeasure. If f is a measurable function such that there exists a sequence
ofZ simple functions (fn )n which pointwise converges to f and the sequences
( fn± dN )n are Cauchy in (cb(X), dH ), then f is BDSm integrable with
E
respect to N .
Notice first that if f = ni=1 ai 1Ei is measurable with non-negative ai , i =
P
Z n
M
1, . . . , n, and N is cb(X)-valued, then f dN = ai N (E ∩ Ei ) for every
E i=1
E ∈ Σ, since for every ∈ x′ X′
the support function is additive with respect
to the Minkowski addition.
In the general case, if a sequence (fn )n of simple functions converges point-
wise to f , then (fn± )n converges to f ± . We consider first f + . (fn+ )n converges
poitwise to f + and for every x′ ∈ X ′ and for every E ∈ Σ
Z   Z 
+ ′ ′ +
fn ds(x , N ) = s(x , fn dN ) .
E n E n
Z 
Since fn+ dN is Cauchy in (cb(X), dH ), by the completeness of the
E n
hyperspace for every E ∈ Σ there exists M + (E) ∈ cb(X) such that
Z 
+ +
lim dH fn dN, M (E) = 0.
n→∞ E
We apply now [19, Lemma 2.3] to fn+ , f + and s(x′ , N ) and we obtain that
f + is integrable with respect to s(x′ , N ) and
Z Z
+ ′
lim fn ds(x , N ) = f + ds(x′ , N ), uniformly with respect to E ∈ Σ.
n→∞ E E
Since
Z  Z
dH fn+ dN, M + (E) = sup s(x , ′
fn+ dN ) − s(x′ , M + (E))
E x′ ∈BX ′ E

we have that Z
′ +
s(x , M (E)) = f + ds(x′ , N ).
E
For the negative part f − we apply the same construction using s(x′ , −N );
in this way we obtain analogously M − (E). Finally, considering M + (E) ⊕
M − (E) we obtain
Z Z
′ + − + ′
s(x , M (E) ⊕ M (E)) = f ds(x , N ) + f − ds(x′ , −N ).
E E
Similarly, if f is a BDSm-integrable function with respect to a dH -multimeasure
N , then there exists a sequence (fn )n of simple functions that is pointwise
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 9
Z 
convergent to f and the sequence fn dN is Cauchy in (cb(X), dH ).
E n
See Remark 5.3 for the proof.
Proposition 2.6. If N is a positive multimeasure, then the BDS m -integral
with respect to N is a sublinear function of its integrands.
Proof. Indeed, assume that f, g are BDS m -integrable with respect to N and
a, b ≥ 0. Then,
 Z  Z Z
′
s x , (af + bg) dN = (af + bg) ds(x , N ) + (af + bg)− ds(−x′ , N )
+ ′
E E E
Z Z
≤ (af + bg ) ds(x , N ) + (af − + bg− ) ds(−x′ , N )
+ + ′
E E
Z Z 
= a f + ds(x′ , N ) + f − ds(−x′ , N )
Z E Z
E

+ ′ − ′
+ b g ds(x , N ) + g ds(−x , N )
E E
 Z   Z 
′ ′
= as x , f dN + bs x , g dN .
E E

The following lemma is essential for our further investigation:
Lemma 2.7. Let M, N : Σ → c(X)Zbe two multimeasures possessing control
measure µ and such that M (E) = θ dN , for every E ∈ Σ. Then M is
E
pointless if and only if N is pointless. Equivalently, M is a vector measure
if and only if N is a vector measure.
Proof. Assume that N (E) = {κ(E)} for every E ∈ Σ and κ : Σ → X is a
vector measure. Then, we have for each E ∈ Σ and each x′ ∈ X ′
Z Z Z Z
′ + ′ − ′ + ′
s(x , M (E)) = θ ds(x , N ) + θ ds(x , −N ) = θ dhx , κi + θ − dhx′ , −κi
E E E E
Z Z Z
+ ′ − ′
(6)= θ dhx , κi − θ dhx , κi = θ dhx′ , κi .
E E E
Z
By our assumption the integral θ dhx′ , κi exists and has to be finite. If
Z E
not, then the equality θ dhx′ , κi = +∞ yields s(−x′ , M (E)) = −∞, what
E
is impossible. Thus, M has only bounded sets as its values. In such a case
the expression on the right hand side of (6) is a linear function on X ′ . Hence
the same holds true for x′ −→ s(x′ , M (E)). But that means that M is a
vector measure.
A similar situation takes place if M (E) := {ν(E)} , E ∈ Σ, where ν is a
vector measure. By the assumption, we have
10 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

Z Z
′ ′ ′ + ′
∀ E ∈ Σ, ∀x ∈X, hx , ν(E)i = θ ds(x , N ) + θ − ds(x′ , −N ) .
E E
Let A := {ω : θ(ω) > 0}. By the classical Radon-Nikodym theorem there
exists a function fx′ such that
Z
′ ′ ′
(7) ∀ E ∈ ΣA , ∀ x ∈ X , s(x , N (E)) = fx′ dµ .
E
If E ∈ ΣA , then
Z
′ ′ ′
∀ E ∈ ΣA , ∀x ∈X, hx , ν(E)i = θ + ds(x′ , N ) .
E
Since for every E ∈ ΣAZ, the function x′ −→ hx′ , ν(E)i is linear, the
same holds true for x′ −→ θ + ds(x′ , N ). Consequently, if a, b ∈ R and
E
x′ , y ′ ∈ X ′ , then
Z Z
+
θ fax +by dµ =
′ ′ θ + ds(ax′ + by ′ , N ) =
E
ZE Z
+ ′ ′
= θ d[a · s(x , N ) + b · s(y , N )] = θ + [afx′ + bfy′ ] dµ .
E E
As E ∈ ΣA is arbitrary and θ + |A > 0, we obtain the equality
fax′ +by′ = afx′ + bfy′ µ − a.e.
Then x′ −→ s(x′ , N (E)) is linear by (7) and this proves that N is a vector
measure on ΣA . Similarly for Ac . 
Proposition 2.8. Let M, N : Σ → c(X) be two multimeasures and assume
that N is pointless. If θ : Ω → R is a measurable function such that for each
E ∈ Σ and x′ ∈ X ′
Z
(8) s(x′ , M (E)) = θ ds(x′ , N ) ,
E
then θ is non negative N -almost everywhere.
Proof. Let θ = θ + − θ − and let H := {ω ∈ Ω : θ(ω)− > 0} ∈ Σ. Then, we
obtain for every E ∈ ΣH the equality
Z
′
s(x , M (E)) = −θ − ds(x′ , N ) .
E
It is enough to prove that N (H) = {0}. We suppose, by contradiction,
that N (H) 6= {0}. Then x′ −→ s(x′ , M (E)) is sublinear for each E ∈
Σ
R H −and also x′ −→ −s(x′ , M (E)) is sublinear, because −s(x′ , M (E)) =
′ ′ ′
E θ ds(x , N ). Hence x → s(x , M (E)) is linear. Therefore
0 = s(0, M (E)) = s(x′ − x′ , M (E)) = s(x′ , M (E)) + s(−x′ , M (E))
and so s(−x′ , M (E)) = −s(x′ , M (E)) 6= ±∞, what yields the linearity of
x′ −→ s(x′ , M (E)). But that is possible only if M (E) is one point set.
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 11

This however forces N to be a vector measure on ΣH , what contradicts the

pointlessness of N . 
Remark 2.9. It follows from Proposition 2.8 that an integral defined by
the equality (8) does not present the proper approach to integrability with
respect to a pointless multimeasure, since only non-negative functions could
be integrable.
If in Proposition 2.8 N restricted to an element F ∈ / N (N ) is a vector
measure, then M restricted to F is a vector measure (see Lemma 2.7) and
θ|F is not necessarily non-negative.

3. Control measures
Our aim is to determine when a multimeasure M can be seen as an integral
of a scalar function with respect to a given multimeasure N . We obtain our
results under the assumption of the existence of control measures for the
considered multimeasures. In the case of dH -multimeasures control measures
always exist (see [1]), but in the case of an arbitrary multimeasure this is not
obvious. The subsequent theorem describes completely the class of Banach
spaces where every multimeasure is a dH -multimeasure.
Theorem 3.1. Every cb(X)-valued multimeasure is a dH -multimeasure if
and only if X does not contain any isomorphic copy of c0 .
Proof. If c0 * X isomorphically, then the assertion is proved in [8, Proposi-
tion 4.1]. If c0 can be isomorphically embedded into X, then [25, Example
3.6 and 3.8] are two examples of cb(c0 )-valued multimeasures which are not
dH -multimeasures. 

The proofs below provide a characterisation of multimeasures possessing

a control measure in the language of the countable chain condition. The
proofs are related to those given in [23], where vector measures were under
consideration.
Definition 3.2. A multimeasure M : Σ → c(X) satisfies the countable
chain condition (ccc) if each family of pairwise disjoint not M -null sets is
at most countable.
Lemma 3.3. Assume that M, N : Σ → c(X) are two multimeasures such
that M ≪ N . If N satisfies (ccc), then for every A ∈ Σ \ N (M ) there exists
B ∈ Σ \ N (M ) such that B ⊂ A and N ≪ M on ΣB .
Proof. Assume that there is a set A ∈ Σ \ N (M ) such that for every B ∈
ΣA \N (M ) there exists D ∈ [N (M )∩B]\N (N ). The lemma of Kuratowski-
Zorn and (ccc) give the existence of at most countable maximal family {Dn }n
of pairwise disjoint sets Dn ∈
S [N (M ) ∩ A] \ N (N ).
OneScan easily check that n Dn ∈ N (M ). Since A ∈ / N (M ), we have
A \ n Dn ∈ / N (M ), but this contradicts the maximality of {Dn }n . 
12 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

Lemma 3.4. If N : Σ → c(X) is a multimeasure satisfying (ccc), then there

exists at most countable family {x′n : n ∈ N} ⊂ X ′ satisfying the equality
\ \
(9) N [s(x′n , N )] = N [s(x′ , N )] .
n x′ ∈X ′

Proof. It follows from Lemma 3.3 that for every x′ ∈ X ′ there exists Dx′ ∈
Σ \N [s(x′ , N )] such that N ≪ s(x′ , N ) on ΣDx′ . The lemma of Kuratowski-
Zorn and (ccc) guarantee existence of at most countable maximal family
{Dn }n ofSdisjoint sets Dn corresponding to measures s(x′n , N ).
Let D = n Dn . If A ∈ Σ and A ∩ D = ∅, then clearly A ∈ N [s(x′ , N )], for
every x′ ∈ BX ′ .
Let now A be an s(x′n , N )-null set, for every n. We have A ∩ Dn ∈ N (N )
for all n and so A ∩ Dn ∈ N [s(x′ , N )] for every x′ . Consequently, A ∩ D ∈
N [s(x′ , N )] for every x′ . Hence,
A = (A \ D) ∪ (A ∩ D) ∈ N [s(x′ , N )] .
That proves (9). 
Theorem 3.5. A multimeasure N : Σ → c(X) has a finite control measure
if and only if it satisfies (ccc). Then, there exists a control measure that is
equivalent to N .
Proof. It is obvious that the existence of a control measure yields (ccc) of N .
So assume that N satisfies (ccc). For each x′ ∈ X ′ let νx′ : Σ → [0, +∞) be
a measure equivalent to s(x′ , N ). By Lemma 3.4 there exist x′n ∈ X ′ , n ∈ N,
such that
\ \ \
N (N ) = N [s(x′ , N )] = N [s(x′n , N )] = N (νx′n ) .
x′ ∈X ′ n n
The measure
∞
X 1 νx′n (E)
(10) µ(E) :=
2n 1 + νx′n (Ω)
n=1
is the required control measure for N that is equivalent to N . 
Theorem 3.6. If X ′ is weak′ -separable, then every multimeasure M : Σ →
c(X) satisfies (ccc).
Proof. Let {x′n : n ∈ N} be weak′ -dense in X ′ . Suppose there exists a
multimeasure M : Σ → c(X) and an uncountable family {Bα ∈ Σ : α ∈ A}
of pairwise disjoint sets with M (Bα ) 6= {0}. Without loss of generality we
may assume that the family is ordered by the ordinals less than ω1 . Due
to the countability of the weak′ -dense set, there exists β < ω1 such that
s(x′n , M (Bα )) = 0, for every α > β and every n ∈ N. Hence, if α > β and
x ∈ Bα , then hx′n , xi ≤ 0 for every n. It follows that hx′ , xi ≤ 0 for every
x′ ∈ X ′ , if x ∈ Bα and α > β. This is of course impossible for x 6= 0. 
The following recent result describes Banach spaces with all cb(X)-multimeasures
possessing control measures.
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 13

Theorem 3.7. [28] If X is a Banach space not containing any isomorphic

copy of c0 (ω1 ), then each multimeasure M : Σ → cb(X) admits a control
measure.

4. Arbitrary multimeasures
Now we start by examining the problem when a multimeasure M can
be represented as an integral of a scalar function with respect to a given
multimeasure N , for suitable multisubmeasures this problem was also faced
in [5] for the Gould integral.
In the case of vector measures with values in a locally convex space a Radon-
Nikodým theorem for the Bartle-Dunford-Schwartz integral was obtained by
Musial in [22]. If Y is a locally convex space and ν, κ : Σ → Y are vector
measures, then the following definitions were formulated in [22]:
usac): ν is uniformly scalarly absolutely continuous (usac) with respect to
κ, if for each ε > 0 there exists δ > 0 such that for each y ′ ∈ Y ′ and
each E ∈ Σ, the inequality |y ′ κ|(E) < δ yields |y ′ ν|(E) < ε. We
denote it by ν ≪ κ.
usd): ν is uniformly scalarly dominated (usd) by κ, if there exists b ∈ R+
such that ∀ y ′ ∈ Y ′ , ∀ E ∈ Σ it is |y ′ ν|(E) ≤ b |y ′ κ|(E).
sub): ν is subordinated to κ, if there exists d ∈ R+ such that for every
E∈Σ
ν(E) ∈ d aco R(κE ).
We say that a vector measure ν : Σ → Y has locally a property with respect
to a vector measure κ : Σ → Y , if for each E ∈ Σ \ N (κ) there exists F ⊂ E
with F ∈ Σ \ N (κ) such that ν has this property with respect to κ on the
set F .

In order to find proper conditions guarateeing the differentiation of an

arbitrary cb(X)-valued multimeasure M with respect to N , we must adapt
properly the definitions given for vector measures.
Definition 4.1. Given two multimeasures M, N : Σ → cb(X) we say that:
(usac): M is uniformly scalarly absolutely continuous with respect to N
(usac or M ≪ N ), if there exists A ∈ Σ such that
> 0, ∃ δ > 0 : ∀ α, β ∈ R, ∀ x′ , y ′ ∈ X ′ , ∀ E ∈ Σ
∀ ε(11)
|αs(x′ , N ) + βs(y ′ , N )|(E ∩ A) + |αs(x′ , −N ) + βs(y ′ , −N )|(E ∩ Ac ) ≤ δ ⇒
 

⇒ |αs(x′ , M ) + βs(y ′ , M )|(E) ≤ ε;

(usd): M is uniformly scalarly dominated (usd) by a multimeasure N ,
if there exist c ∈ R and A ∈ Σ such that ∀ α, β ∈ R, ∀ x′ , y ′ ∈
X ′, ∀ E ∈ Σ
|αs(x′ , M ) + βs(y ′ , M )|(E) ≤
≤ c|αs(x′ , N ) + βs(y ′ , N )|(E ∩ A) + c|αs(x′ , −N ) + βs(y ′ , −N )|(E ∩ Ac ) ;
14 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

(uss): M is uniformly scalarly subordinated (uss) to N , if there exist

d ∈ R+ and A ∈ Σ such that ∀ α, β ∈ R, ∀ x′ , y ′ ∈ X ′ , ∀ E ∈ Σ

αs(x′ , M (E)) + βs(y ′ , M (E)) ∈

∈ d aco{[αs(x′ , N (F ∩ A)) + βs(y ′ , N (F ∩ A))] : F ∈ ΣE } +
+ d aco{[αs(x′ , −N (F ∩ Ac )) + βs(y ′ , −N (F ∩ Ac ))] : F ∈ ΣE } .

Remark 4.2. If M and N are vector measures then the uniform scalar
absolute continuity and the uniform scalar domination of M with respect to
N coincide respectively with the corresponding definitions given for vector
measures, therefore we use the same notation for the two notions.

In case of uss the definition looks different with respect to sub and so we
decided to change the name. They actually are equivalent if we assume that
X is a quasi-complete locally convex space. This is due to the fact that for
every E ∈ Σ and for every x′ ∈ X ′

(12) aco{hx′ , R(κE )i} = hx′ , aco(R(κE ))i with

hx′ , R(κE )i := {hx′ , wi : w ∈ R(κE )}.

Here the inclusion ⊃ is a consequence of the continuity of x′ ∈ X ′ :

hx′ , aco(R(κE ))i ⊂ hx′ , aco(R(κE ))i = aco{hx′ , R(κE )i} .

While for the reverse inclusion, we can observe that the set aco R(κE ) is
symmetric and weakly compact, due to [21, Theorem IV.6.1].
So we have hx′ , aco R(κE )i = [−α, α]. The inclusion R(κE ) ⊂ aco R(κE )
yields hx′ , R(κE )i ⊂ [−α, α] and so aco hx′ , R(κE )i ⊂ [−α, α].

Proposition 4.3. For arbitrary multimeasures M, N : Σ → cb(X) the pro-

perties usac and usd are equivalent and uss implies each of them.

Proof. usd) ⇒ usac) is obvious.

(usac) ⇒ (usd) Let A ∈ Σ, ε > 0 and δ > 0 be such that |αs(x′ , N ) +
βs(y ′ , N )|(E) < δ implies |αs(x′ , M ) + βs(y ′ , M )|(E) < ε.
If |αs(x′ , N )+βs(y ′ , N )|(E) = 0, then M ≪ N yields |αs(x′ , M )+βs(y ′ , M )|(E) <
ε for every ε > 0 . So usd follows.
Suppose now that |αs(x′ , N ) + βs(y ′ , N )|(E) > 0. Let

x̂ := δx′ [2|αs(x′ , N ) + βs(y ′ , N )|(E)]−1 ŷ := δy ′ [2|αs(x′ , N ) + βs(y ′ , N )|(E)]−1

Then |αs(x̂, N )+βs(ŷ, N )|(E) < δ and consequently |αs(x̂, M )+βs(ŷ, M )|(E) ≤
ε. It follows that, if we take c = 2ε/δ, then

|αs(x′ , M ) + βs(y ′ , M )|(E) ≤ c|αs(x′ , N ) + βs(y ′ , N )|(E) .

In a similar way one obtains the required inequalities for very E ∈ ΣAc .
That proves usd.
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 15

(uss) ⇒ (usd) By definition, for every F ∈ ΣE , we have

αs(x′ , M (E)) + βs(y ′ , M (E)) ≤
≤ sup d aco{[αs(x′ , N (F ∩ A)) + βs(y ′ , N (F ∩ A))]} +
+ sup d aco{[αs(x′ , −N (F ∩ Ac )) + βs(y ′ , −N (F ∩ Ac ))]} ≤
≤ d|αs(x′ , N ) + βs(y ′ , N )|(E ∩ A) +
+ d|αs(x′ , −N ) + βs(y ′ , −N )|(E ∩ Ac ) .
So the assertion follows. 
Theorem 4.4. Let M, N : Σ → cb(X) be two consistent multimeasures
possessing control measures. Then the following are equivalent:
(RNb ): There exists a bounded measurable function θ : Ω → R such that
for every E ∈ Σ we have
Z
(13) M (E) = θ dN ;
E
(4.4.i): M is uniformly scalarly dominated by N ;
(4.4.ii): M is uniformly scalarly absolutely continuous with respect to N ;
(4.4.iii): M is uniformly scalarly subordinated to N .
Proof. Let µ be a finite control measure for N . Without loss of generality,
we may assume that µ is also a control measure for M . Let H ∈ Σ be a set
as in the definition of the consistency for M and N . We divide the proof
into the pointless and vector parts.
(Pointless part): We assume for simplicity that H = Ω.
(4.4.i) ⇒ (RNb ): By the classical Radon-Nikodým theorem, for
every x′ ∈ X ′ , there exist two measurable real functions fx′
and gx′ such that, ∀ E ∈ Σ,
Z Z
′ ′
(14) s(x , M (E)) = fx dµ and s(x , N (E)) =
′ gx′ dµ .
E E
• Let A be the set satisfying the definition of uniform scalar
domination of M with respect to N . If E ∈ ΣA and
α, β ∈ R, then
Z
|αfx′ + βfy′ | dµ = |αs(x′ , M ) + βs(y ′ , M )|(E)
E
≤ c|αs(x′ , N ) + βs(y ′ , N )|(E)
Z
= c |αgx′ + βgy′ | dµ.
E
So |αfx′ + βfy′ | ≤ c|αgx′ + βgy′ | µ − a.e. on A. Hence
(see also [22, Lemma]),

f x′ fy ′
= µ − a.e. on the set {ω ∈ Ω : gx′ (ω)gy′ (ω) 6= 0} ∩ A .
cgx′ cgy′
16 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

According to [31, Theorem 7.35.2] (see also [22, Lemma]

for a short proof) there exists a measurable θ1 : A →
[−1, 1] such that fx′ = c θ1 gx′ µ-a.e. on the set {ω ∈
A : gx′ (ω)gy′ (ω) 6= 0} for each x′ ∈ X ′ separately. The
equality on the set {ω ∈ A : gx′ (ω) = 0} is obvious
(notice that |fx′ | ≤ c|gx′ | a.e. on A). So it follows that

Z
′
s(x , M (E)) = c θ1 ds(x′ , N )
E

for each E ∈ ΣA and x′ ∈ X ′ . We can prove that θ1 ≥ 0

N -a.e. on A, thanks to Proposition 2.8.
• Let now F ∈ ΣAc . For every α, β ∈ R and for every
x′ , y ′ ∈ X ′ we have
Z
|αfx′ + βfy′ | dµ = |αs(x′ , M ) + βs(y ′ , M )|(F )
F
≤ c|αs(x′ , −N ) + βs(y ′ , −N )|(F )
Z
= c |αg−x′ + βg−y′ | dµ.
F

A similar calculation gives the equality

f x′ fy ′
= µ − a.e. on the set {ω ∈ Ac : g−x′ (ω)g−y′ (ω) 6= 0}
cg−x′ cg−y′
and then

Z
′
s(x , M (F )) = c θ2 ds(x′ , −N ) .
F

As before, θ2 : Ac → R is non-negative N -a.e.. Thus, set

θ = cθ1 − cθ2 , for every E ∈ Σ

s(x′ , M (E)) = s(x′ , M (E ∩ A)) + s(x′ , M (E ∩ Ac )) =

Z Z
+ ′
= θ ds(x , N ) + θ − ds(x′ , −N ) ,
E E

what means that θ is a Radon-Nikodým derivative of M

with respect to N .
( RNb ) ⇒ (4.4.iii): We set A := {ω ∈ Ω : θ(ω) ≥ 0}, and d >
supω∈Ω |θ(ω)|. Since θ is BDSm-integrable with respect to N ,
by Definition 2.3 we have for each E ∈ Σ the equality
Z
αs(x′ , M (E ∩ A)) + βs(y ′ , M (E ∩ A)) = θ + d[αs(x′ , N ) + βs(y ′ , N )] .
E∩A
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 17

Let x′ , y ′ , α, β be fixed and let B, C generate the Hahn decom-

position of αs(x′ , N ) + βs(y ′ , N ). Then,
αs(x′ , M (E ∩ A ∩ B)) + βs(y ′ , M (E ∩ A ∩ B))
∈ co [θ + (E ∩ A ∩ B)] · [αs(x′ , N (E ∩ A ∩ B)) + βs(y ′ , N (E ∩ A ∩ B))]
Similarly,
αs(x , M (E ∩ A ∩ C)) + βs(y ′ , M (E ∩ A ∩ C))
′

∈ co [θ + (E ∩ A ∩ C)] · [αs(x′ , N (E ∩ A ∩ C)) + βs(y ′ , N (E ∩ A ∩ C))]

All together yields
αs(x′ , M (E ∩ A)) + βs(y ′ , M (E ∩ A))

∈ co [θ + (E ∩ A)] · [αs(x′ , N (E ∩ A ∩ B)) + βs(y ′ , N (E ∩ A ∩ B))]

+ [αs(x′ , N (E ∩ A ∩ C)) + βs(y ′ , N (E ∩ A ∩ C))]
= co [θ + (E ∩ A)] · [αs(x′ , N (E ∩ A)) + βs(y ′ , N (E ∩ A))]
⊂ d aco {αs(x′ , N (F ∩ A)) + βs(y ′ , N (F ∩ A)) : F ∈ ΣE }.
In the same way one obtains the inclusion
αs(x′ , M (E ∩ Ac )) + βs(y ′ , M (E ∩ Ac ))
⊂ d aco {αs(x′ , −N (F ∩ Ac )) + βs(y ′ , −N (F ∩ Ac )) : F ∈ ΣE }.
The remaining equivalences follow from Proposition 4.3.
(Vector part): We assume now that H c = Ω. In virtue of Remark 4.2, the
equivalence of the four conditions is a consequence of [22, Theorem
1].

Similarly to the vector case we say that a multimeasure M : Σ → cb(X)
has locally a property with respect to a multimeasure N : Σ → cb(X), if
for each E ∈ Σ \ N (N ) there exists F ⊂ E with F ∈ Σ \ N (N ) such that
M has the property with respect to N on the set F .

Using the local properties we obtain:

Theorem 4.5. Let M, N : Σ → c(X) be two σ-bounded consistent multi-
measures possessing control measures. Then the following are equivalent:
(RN): There exists a measurable function θ : Ω → R such that
Z
∀E∈Σ M (E) = θ dN ;
E
(4.5.i): M is locally uniformly scalarly dominated by N ;
(4.5.ii): M is locally uniformly scalarly absolutely continuous with respect
to N ;
(4.5.iii): M is locally uniformly scalarly subordinated to N .
18 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI
S
Proof. Assume that Ω = n Ωn ∪ B, where Ωn -s are pairwise disjoint, B ∈
N (µ) and the restriction of M and N to each Ωn is cb(X)-valued. First we
assume that M, N are cb(X)-valued on Σ.
Analogously to Theorem 4.4 we have to divide the proof into the pointless
and vector parts. The vector part follows immediately from [22, Theorem
2], so we prove here only the pointless part.
• The equivalences (4.5.i) ⇔ (4.5.ii) ⇔ (4.5.iii) follow from the
corresponding equivalences in Theorem 4.4.
• Also (RN) ⇒ (4.5.i) is obvious (it is enough to take for each
E ∈ Σ \ N (N ) a set F ⊂ E such that F ∈ Σ \ N (N ) and θ is
bounded on F ).
• Now we are going to prove that (4.5.i) ⇒ (RN).
Let µ be a control measure for N . By definition there exist E ∈
Σ \ N (µ) such that M is scalarly dominated on E by N . We denote
by

H1 := {E ∈ Σ \ N (µ) : ∃ AE ∈ ΣE ∀ F ∈ ΣE , ∀ x′ , y ′ ∈ X ′ , ∀ α, β ∈ R,
|αs(x′ , M ) + βs(y ′ , M )|(F )
≤ |αs(x′ , N ) + βs(y ′ , N )|(F ∩ AE ) + |αs(x′ , −N ) + βs(y ′ , −N )|(F \ AE )} .

By the completeness of the algebra Σ/N (µ) there is E1 ∈ Σ such

that its equivalence class is the least upper bound of H1 in Σ/N (µ).
If H1 is empty we choose E1 = ∅.
Then we consider the class H2 of all sets E ∈ ΣΩ\E1 \N (µ) in which
we have the scalar domination with respect to c = 2 and we choose
analogously E2 .
After n steps we have already sets in this way E1 , . . . , En and AE1 ⊂
E1 , . . . , AEn ⊂ En such that for each k ≤ n the inequality

|αs(x′ , M ) + βs(y ′ , M )|(F )

≤ k |αs(x′ , N ) + βs(y ′ , N )|(F ∩ AEk ) + |αs(y ′ , −N ) + βs(x′ , −N )|(F \ AEk )


holds true for all F ∈ ΣEk , for all α, β ∈ R and for all x′ , y ′ ∈ X ′ .
Then we construct En+1 such that its equivalence class is the least
upper bound of

Hn+1 := {E ∈ ΣΩ\Snk=1 Ek \ N (λ) : ∃ AE ∈ ΣE ∀ F ∈ ΣE , ∀ x′ , y ′ ∈ X ′ , ∀ α, β ∈ R

|αs(x′ , M ) + βs(y ′ , M )|(F ) ≤ (n + 1)|αs(x′ , N ) + βs(y ′ , N )|(F ∩ AE )
+ (n + 1)|αs(x′ , −N ) + βs(y ′ , −N )|(F \ AE )} .

in Σ/N (µ). A few first sets Ek may be of measure zero but then
we meet the first set Ek of positive measure. In this way we obtain
a sequence (possibly finite) (En ) of sets with En ∈ Hn .
Without loss of generality we may assume that the sets cover all Ω.
Now we apply Theorem 4.4 to each set En and we get a bounded
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 19

measurable function θn such that

(15) Z Z
′ ′ ′
∀ E ∈ ΣE n , ∀ x ∈ X , s(x , M (E)) = θn+ ds(x′ , N )+ θn− ds(x′ , −N ) .
E E
P∞ ′ ′ +
Let
P∞ θ = +
P∞let −x ∈ X be fixed. Then, θ =
n=1 θn χEn and
−
n=1 θn χEn and θ = n=1 θn χEn . By (15) we have for every
E∈Σ
∞
X
s(x′ , M (E)) = s(x′ , M (E ∩ En )) =
n=1
X∞ Z Z 
(16) = θn+ ds(x′ , N ) + θn− ds(x′ , −N ) .
n=1 E∩En E∩En

Let Ω + := {ω ∈ Ω : θ(ω) ≥ 0}, Ω − := Ω \Ω + . Let P1′ , P2′ , Q′1 , Q′2 ∈

Σ be two Hahn decomposition for s(x′ , N ), s(x′ , −N ) respectively.
We have
∞
X
s(x′ , M (E ∩ P1′ ∩ Ω + )) = s(x′ , M (E ∩ En ∩ P1′ ∩ Ω + )) =
n=1
X∞ Z
= θn+ ds(x′ , N )
n=1 E∩En ∩P1′ ∩Ω +

X∞
s(x′ , M (E ∩ P2′ ∩ Ω + )) = s(x′ , M (E ∩ En ∩ P2′ ∩ Ω + )) =
n=1
X∞ Z
= θn+ ds(x′ , N )
n=1 E∩En ∩P2′ ∩Ω +

∞
X
′
s(x , M (E ∩ Q′1 −
∩ Ω )) = s(x′ , M (E ∩ En ∩ Q′1 ∩ Ω − )) =
n=1
X∞ Z
= θn− ds(x′ , −N )
′
n=1 E∩En ∩P1 ∩Ω
−

X∞
s(x′ , M (E ∩ Q′2 ∩ Ω − )) = ′
s(x , M (E ∩ En ∩ Q′2 ∩ Ω − )) =
n=1
X∞ Z
= θn− ds(x′ , −N ).
n=1 E∩En ∩P2′ ∩Ω −

Each of the series closing the above equalities is convergent with all
its terms of the same sign. Hence, they are absolutely convergent.
20 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

It follows that for all E ∈ Σ

X∞ Z Z
′ ′ + ′
s(x , M (E)) = s(x , M (E ∩ En )) = θ ds(x , N ) + θ + ds(x′ , N ) +
n=1 E∩P1′ E∩P2′
Z Z
+ θ − ds(x′ , −N ) + θ − ds(x′ , −N ) =
E∩Q′1 E∩Q′2
Z Z
= θ + ds(x′ , N ) + θ − ds(x′ , −N ).
E E
Let us consider now the general case. We know already that for each
n ∈ N there exists a measurable function ξn : Ωn → R such that
(17) Z Z
∀ E ∈ ΣΩn ∀ x′ ∈ X ′ s(x′ , M (E)) = ξn+ ds(x′ , N ) + ξn− ds(x′ , −N ) .
E E
Then, we follow the proof presented after formula (15). We have to re-
member only that we have always s(x′ , M (E)) > −∞ and so each series
appearing in the proof is either divergent to +∞ or convergent. 
Remark 4.6. It follows from Lemma 2.7 that some introductory assump-
tions in Theorem 4.4 concerning M and N are necessary. If M is pointless
but N is not, then M cannot be represented by a BDSm -integral with re-
spect to N . To construct an example let X be a quasi-complete locally
convex space and N := ν : Σ → X be a non-atomic vector measure.

If M is defined by the formula M (E) := co R(νE ) (or M (E) := aco R(νE )),
then aco R(νE ) is a weakly compact set (see [21, Theorem IV.6.1]) and the
conditions uss, usd and usac are fulfilled by M and N . Indeed the formula
(12) in Remark 4.2, shows that
αs(x′ , M (E)) + βs(x′ , M (E)) ∈ d aco{hx′ , R(νE )i} ,
whenever E ∈ Σ and x′ ∈ X ′ are arbitrary. It remains to prove that
M (E) := co R(νE )
is a multimeasure. M is clearly finitely additive: if A, B are disjoint, then
M (A∪B) = co R(νA∪B ) = co (R(νA )+R(νB )) = co R(νA )⊕co R(νB ) = M (A)⊕M (B).
It follows that for every x′ ∈ X ′ , s(x′ , M ) : Σ → R is finitely additive. [3,
Proposition 3.8] yields the countable additivity of each s(x′ , M ) and so M
is a multimeasure.
Remark 4.7. A particular example of a quasi-complete locally convex space
is a conjugate Banach space endowed with the weak∗ topology σ(X ′ , X). Let
cw∗ k(X ′ ) denote the family of all non-empty weak∗ -compact and convex sub-
sets of X ′ . M : Σ → cw∗ k(X ′ ) is called a weak∗ -multimeasure if s(x, M (·)) is
a measure, for every x ∈ X. It is proved in [26, Theorem 3.4] that X does not
contain any isomorphic copy of ℓ∞ if and only if each weak∗ -multimeasure
M : Σ → cw∗ k(X ′ ) is a dH -multimeasure. The Radon-Nikodým theorem
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 21

for two weak∗ -multimeasures formulates exactly as Theorem 4.5, one should
only remember that now the conjugate to (X ′ , σ(X ′ , X)) is the space X
itself.

5. dH -multimeasures and their Rådström embeddings

Throughout this section we assume that X is a Banach space and we
consider now the case of dH -multimeasures. Let us recall that if a Banach
space X does not contain any isomorphic copy of c0 , then each cb(X)-valued
multimeasure is a dH -multimeasure (see [6, 8]). Examples of cb(c0 )-valued
measures that are not dH -measures can be found in [25, Example 3.6, Ex-
ample 3.8]. When we write nm=1 am x′m ∈ spanBX ′ , we always mean that
P
x′m ∈ BX ′ and am ∈ R, for every m ≤ n.

Let now M, N : Σ → cb(X) be two dH -multimeasures. We remember

that ks(·, M (E))k∞ = supx′ ∈BX ′ |s(x′ , M (E))| and, according to (1), we set
Y := ℓ∞ (BX ′ ). By [22, Theorem 1], for ν := j ◦ M, κ = j ◦ N we have
Theorem 5.1. If M, N : Σ → cb(X) are two dH -multimeasures, then the
following are equivalent:
RNj ): There exists a bounded measurable function (measurable function)
θ : Ω → R such that for all E ∈ Σ and y ′ ∈ ℓ′∞ (BX ′ )
Z
hy ′ , j ◦ M (E)i = θ dhy ′ , j ◦ N i ;
E

(5.1.i): j ◦ M is (locally) uniformly scalarly absolutely continuous with re-

spect to j ◦ N ;
(5.1.ii): j ◦ M is (locally) uniformly scalarly dominated by j ◦ N ;
(5.1.iii): j ◦ M is (locally) subordinated to j ◦ N .
We observe that even if M and N are dH -multimeasures, the differenti-
ation of M with respect to N is in general not equivalent to differentiation
of j ◦ M with respect to j ◦ N . If θ is the Radon-NikodýmZ derivative of M
with respect to N , the representation j ◦ M (E) = (BDS) θ d(j ◦ N ) for
E
every E ∈ Σ may fail (see the subsequent Example 6.4). In fact one has to
take into account also integration with respect to the measure j ◦ (−N ). If
N is a vector measure, then j ◦ (−N ) = −j ◦ N and formally the measure
j ◦ (−N ) is absent in the calculations and the integral with respect to N
coincides with its BDS-integral.

In the case of multimeasures that are not vector measures, we have in

general j ◦ (−N ) 6= −j ◦ N and the presence of the measure j ◦ (−N )
becomes visible. Next theorem shows the shape of j ◦ N provided M has
the Radon-Nikodým derivative (in the sense of the integral investigated in
this paper) with respect to N .
22 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

Theorem 5.2. Let M, N : Σ → cb(X) be two dZH -multimeasures and θ :

Ω → R be a measurable function. Then M (E) = θ dN for every E ∈ Σ
E
if and only if for all E ∈ Σ and for all y ′ ∈ ℓ′∞ (BX ′ )
Z Z
′ + ′
(18) hy , j ◦ M (E)i = θ dhy , j ◦ N i + θ − dhy ′ , j ◦ (−N )i .
E E
Equivalently,
Z Z
+
j ◦ M (E) = (BDS) θ d(j ◦ N ) + (BDS) θ − d(j ◦ (−N )) .
E E

Proof. ⇒ : (5.2.A):
Z We assume first that θ is bounded. Let M (E) =
θ dN for every E ∈ Σ. According to (5) we have then
E
Z Z
hex′ , j ◦ M (E)i = s(x′ , M (E)) = θ + ds(x′ , N ) + θ − ds(−x′ , N )
E E
Z Z
= θ + dhex′ , j ◦ N i + θ − dhe−x′ , j ◦ N i =
E E
Z Z
+
= θ dhex′ , j ◦ N i + θ − dhex′ , j ◦ (−N )i.
E E
Since θ is bounded, it is j ◦ N integrable (as a BDS-integral)
and so there are measures κ1 , κ2 : Σ → ℓ∞ (BX ′ ) such that
Z
′
hy , κ1 (A)i = θ + dhy ′ , j ◦ N i whenever y ′ ∈ ℓ′∞ (BX ′ )
E
and
Z
′
hy , κ2 (E)i = θ − dhy ′ , j ◦ (−N )i whenever y ′ ∈ ℓ′∞ (BX ′ ).
E
Hence, we obtain the equality
hex′ , j ◦ M (E)i = hex′ , κ1 (E)i + hex′ , κ2 (E)i.
But {ex′ : x′ ∈ BX ′ } is norming and the measures j ◦ M, κ1
and κ2 have weakly relatively compact ranges. It follows that
∀ y ′ ∈ ℓ′∞ (BX ′ ), ∀ E ∈ Σ
Z Z
(19) hy ′ , j ◦ M (E)i = θ + dhy ′ , j ◦ N i + θ − dhy ′ , j ◦ (−N )i.
E E

Assume now that the equality (19) is fulfilled. Then, ∀ x′ ∈

BX ′ , ∀ E ∈ Σ
Z Z
s(x′ , M (E)) = hex′ , j ◦ M (E)i = θ + dhex′ , j ◦ N i + θ − dhex′ , j ◦ (−N )i
Z ZE E
Z
+ ′
= θ ds(x , N ) + θ ds(x , −N ) = s(x′ ,
− ′
θ dN ) .
E E E
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 23
Z
(5.2.B): Suppose now that θ ≥ 0 is arbitrary and M (E) = θ dN
E
for every E ∈ Σ.
That means that for every x′ ∈ X ′ and E ∈ Σ we have
Z
′
s(x , M (E)) = θ ds(x′ , N ) .
E
P
θ can be represented as θ = n θχEn , where the sets En
are pairwise disjoint and each θχEn is bounded. According
to (5.2.A) part, we have for all y ′ ∈ ℓ′∞ (BX ′ )
Z
′
hy , j ◦ M (E ∩ En )i = θ dhy ′ , j ◦ N i , n ∈ N .
E∩En
Since for each y′ ∈ ℓ′∞ (BX ′ )the set
P function hy ′ , j ◦ M i is
′
a measure, we obtain the equality n hy , j ◦ M (E ∩ En )i =
′ ′ ′
hy , j ◦ M (E)i. Since for each y ∈ ℓ∞ (BX ′ ) the set function
hy ′ , j ◦ Ni is a measure, we have also
XZ Z
′
θ dhy , j ◦ N i = θ dhy ′ , j ◦ N i .
n E∩En E
Z
(5.2.C): θ is arbitrary and M (E) = θ dN for every E ∈ Σ.
E
The proof is obvious.
⇐: In (18) one should substitute ex′ instead of y ′ .

Remark 5.3. Let f be a BDSm -integrable function and assume that N is
cb(X)-valued dH -multimeasure. Then there exists a sequence (f
Z n )n of simple

functions that is pointwise convergent to f and the sequence fn dN
E n
is Cauchy in (cb(X), dRH ).
Indeed, let M (E) = E f dN , E ∈ Σ. According to Theorem 5.2 f + and
f − are j ◦ N integrable as (BDS) integrals) and we have for all E ∈ Σ, with
E ⊂ supp f + Z
j ◦ M (E) = (BDS) f + d (j ◦ N ) .
E
In virtue of [1, Definition 2.5] there exists a sequence of pointwise conver-
gent + hn : supp f + → [0, ∞) such that the sequence
 toZf simple functions
(BDS) hn d (j ◦ N ) is Cauchy in ℓ∞ (BX ′ ). It follows from Theorem
E n
Z 
5.2 that the sequence hn dN is Cauchy in dH . We repeat the pro-
E n
cedure with f − obtaining a sequence (gn )n of simple functions. Setting
fn := hn − gn we find the required sequence.

By Theorem 5.2 and Proposition 2.8 we get

24 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

Corollary 5.4. Let M, N : Σ → cb(X) be two consistent dH -multimeasures.

Moreover let θ : Ω → R be a measurable function.
Z
• If j ◦ M (E) = (BDS) θ d(j ◦ N ) for every E ∈ Σ, then θ ≥
E
0Z N -a.e. and for all E ∈ Σ and all x′ ∈ X ′ s(x′ , M (E)) =
θ ds(x′ , N ).
E
Z
• And conversely, if θ is non-negative and s(x′ , M (E)) = θ ds(x′ , N )
Z E
for every E ∈ Σ, then j ◦ M (E) = (BDS) θ d(j ◦ N ) for every
E
E ∈ Σ.
It is our aim to obtain a Radon-Nikodým theorem for Rådström embed-
dings of multimeasures in terms of the multimeasures themselves, not their
Rådström embeddings. To achieve it we modify the definitions of usac, usd
and uss in the following way:
Definition 5.5. Given two multimeasures M, N : Σ → cb(X) we say that:
(s-usac): A multimeasure M : Σ → cb(X) is strongly uniformly scalarly
absolutely continuous (s-usac) with respect to a multimeasure N :
Pn→ cb(X),
Σ
′
if for each ε > 0 there exists δ P
> 0 such that for each
n ′
m=1 am xm ∈ spanBX and each E ∈ Σ, if |
′
m=1 am s(xm , N )|(E) ≤
δ, then
Xn
am s(x′m , M ) (E) ≤ ε.
m=1
We denote it by M ≪s N .
(s-usd): A multimeasure M : Σ → cb(X) is strongly uniformly scalarly
dominated (s-usd) by a multimeasure N : Σ → cb(X), if there
exists a positive d ∈ R such that for every E ∈ Σ and every
P n ′
m=1 am xm ∈ span BX ′ , one has
n
X n
X
am s(x′m , M ) (E) ≤ d am s(x′m , N ) (E) .
m=1 m=1

(s-uss): A multimeasure M is strongly uniformly scalarly subordinated to

N (s-uss), Pif there exists a positive d ∈ R such that for every E ∈ Σ
and every nm=1 am x′m ∈ span BX ′ , one has
Xn X n 
′ ′
(20) am s(xm , M (E)) ∈ d aco am s(xm , N (F )) : F ∈ ΣE .
m=1 m=1

Using these strong versions of uniform scalar absolute continuity, uniform

scalar domination and uniform scalar subordination we are able to prove the
following result:
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 25

Theorem 5.6. Let M, N : Σ → cb(X) be two consistent dH -multimeasures.

Then the following are equivalent
(RNj ): There exists a bounded measurable function (measurable function)
θ : Ω → R such that for all E ∈ Σ and y ′ ∈ ℓ′∞ (BX ′ )
Z
′
hy , j ◦ M (E)i = θ dhy ′ , j ◦ N i ;
E
(5.6.j): M is (locally) strongly uniformly scalarly dominated by N ;
(5.6.jj): M is (locally) strongly uniformly scalarly absolutely continuous
with respect to N ;
(5.6.jjj): M is (locally) strongly uniformly scalarly subordinated to N .
Proof. (RNj ) ⇒ (5.6.jjj): By Theorem 5.1 the condition (RNj ) implies
that j ◦ M is s-uss to j ◦ N . Assume that d > 0 is such that
j ◦ M (E) ∈ d aco{j ◦ N (F ) : F ∈ ΣE } ∈ cb(ℓ∞ (BX ′ )) for every E ∈ Σ .
This means that for every y ′ ∈ ℓ′∞ (BX ′ ), we have
hy ′ , j ◦ M (E)i ≤ d sup{hy ′ , zi : z ∈ aco{j ◦ N (F ) : F ∈ ΣE } }
= d sup s(y ′ , aco{j ◦ N (F ) : F ∈ ΣE })
≤ d max{s(y ′ , co{j ◦ N (F ) : F ∈ ΣE }), s(y ′ , co{−j ◦ N (F ) : F ∈ ΣE })}
= d max{s(y ′ , {j ◦ N (F ) : F ∈ ΣE }), s(y ′ , {−j ◦ N (F ) : F ∈ ΣE })}.
So, if {x′1 , . . . , x′n } ⊂ X ′ and ai ∈ R, i = 1, . . . , n, then
n
X X n n
X 
ai s(x′i , M (E)) = ai hex′i , j ◦ M (E)i = ai ex′i , j ◦ M (E)
i=1 i=1 i=1
n
 X 
≤ d max s ai exi , {j ◦ N (F ) : F ∈ ΣE } ,
′

i=1
n
X 
s ai ex′i , {−j ◦ N (F ) : F ∈ ΣE }
i=1
 n
X  n
X 
= d max sup ai ex′i , j ◦ N (F ) , sup ai ex′i , −j ◦ N (F )
F ∈ΣE i=1 F ∈ΣE i=1
 n
X   Xn 
= d max sup ai hex′i , j ◦ N (F )i , sup − ai hex′i , j ◦ N (F )i
F ∈ΣE i=1 F ∈ΣE i=1
 Xn   Xn 
′ ′
= d max sup ai s(xi , N (F )) , sup − ai s(xi , N (F ))
F ∈ΣE i=1 F ∈ΣE i=1
 Xn  n
X 
′ ′
= d max sup ai s(xi , N (F )) , − inf ai s(xi , N (F ))
F ∈ΣE i=1 F ∈ΣE
i=1
Xn 
= d sup aco ai s(x′i , N (F )) : F ∈ ΣE .
i=1
26 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

Since we have also

Xn n
X 
′ ′
− ai s(xi , M (E)) ≤ d sup aco ai s(xi , N (F )) : F ∈ ΣE ,
i=1 i=1
we deduce that
Xn n
X 
′ ′
ai s(xi , M (E)) ∈ d aco ai s(xi , N (F )) : F ∈ ΣE .
i=1 i=1
(5.6.jjj) ⇒ (5.6.j): By the assumption
Xn n
X 
′ ′
± ai s(xi , M (E)) ≤ d sup aco ai s(xi , N (F )) : F ∈ ΣE
i=1 i=1
n
X 
≤ d sup aco ai s(x′i , N ) (F ) : F ∈ ΣE
i=1
n
X
≤ d ai s(x′i , N ) (E) .
i=1
Hence, for every set E ∈ Σ we have
X n n
X
ai s(x′i , M )(E) ≤ d ai s(x′i , N ) (E)
i=1 i=1
and the standard calculation gives the required inequality for the
variations.
(5.6.j) ⇒ (5.6.jj): is obvious.
(5.6.jj) ⇒ (RNj ): AssumeP that for each ε > 0 there exists ε/2 > δ > 0
such that for each nm=1 am x′m ∈ span BX ′ and each E ∈ Σ, we
have
Xn Xn
′
am s(xm , N ) (E) ≤ δ, ⇒ am s(x′m , M ) (E) ≤ ε/2.
m=1 m=1
It follows that
X n  n
X 
(21) am ex′m , j ◦ N (E) ≤ δ ⇒ am ex′m , j ◦ M (E) ≤ ε/2 .
m=1 m=1
Let E ∈ Σ and y′∈ be such that |hy ′ , j ◦ N i|(E) ≤ δ/2.
ℓ′∞ (BX ′ )
+ −
Then, let Ey′ ∪ Ey′ = E be the Hahn decomposition of E with
respect to hy ′ , j ◦ N i.
Since {ex′ : x′ ∈ BX ′ } is norming for ℓ∞ (BX ′ ) there exists z ′ ∈
span{ex′ : x′ ∈ BX ′ } such that
|hy ′ − z ′ , j ◦ N (Ey+′ )i| < δ/4 and |hy ′ − z ′ , j ◦ N (Ey−′ )i| < δ/4
and
|hy ′ − z ′ , j ◦ M (Ey+′ )i| < δ/4 and |hy ′ − z ′ , j ◦ M (Ey−′ )i| < δ/4.
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 27

We have then

|hz ′ , j ◦ N i(E)| = |hz ′ , j ◦ N (Ey+′ )i + hz ′ , j ◦ N (Ey−′ )i|

= [hz ′ , j ◦ N (Ey+′ )i − hy ′ , j ◦ N (Ey+′ )i]
+ [hz ′ , j ◦ N (Ey−′ )i − hy ′ , j ◦ N (Ey−′ )i]
+ [hy ′ , j ◦ N (Ey+′ )i + hy ′ , j ◦ N (Ey−′ )i] < δ.

By the assumption |hz ′ , j ◦ M i|(E) ≤ ε/2. Now we follow the re-

verse way:

|hy ′ , j ◦ M i|(E) = hy ′ , j ◦ M (Ey+′ )i + hy ′ , j ◦ M (Ey−′ )i

= [hy ′ , j ◦ M (Ey+′ )i − hz ′ , j ◦ M (Ey+′ )i]
+ [hy ′ , j ◦ M (Ey−′ )i − hz ′ , j ◦ M (Ey−′ )i]
+ [hz ′ , j ◦ M (Ey+′ )i + hz ′ , j ◦ M (Ey−′ )i]
< δ + ε/2 < ε.

That means that j ◦ M is s-usac with respect to j ◦ N and so, in

virtue of Theorem 5.1 j ◦ M has the derivative.

The proof of the local versions is analogous to that given in Theorem 4.5
where the construction is given for the general case of an arbitrary multi-
measure. 

Remark 5.7. Theorems 5.1 and 5.6 allow us to observe that j ◦M ≪ j ◦ N

if and only if M is s-usac with respect to N . Analogously j ◦ M is usd by
j ◦ N if and only if M is s-usd by N .

Remark 5.8. It turns out that the condition (RNj ) implies the following
one (that coincides with the sub condition in case of vector measures):
(sub): ∀ E ∈ Σ M (E) ⊂ d aco [ R(NE ) ∪ R(−NE )] .
In fact, by Theorem 5.1 j ◦ M is subordinated to j ◦ N . Assume that
d > 0 is such that, for every E ∈ Σ,

j ◦ M (E) ∈ d aco{j ◦ N (F ) : F ∈ ΣE } ∈ cb(ℓ∞ (BX ′ )).

This means that for every y ′ ∈ ℓ′∞ (BX ′ ), we have

hy ′ , j ◦ M (E)i ≤ d s(y ′ , aco{j ◦ N (F ) : F ∈ ΣE }) = d s(y ′ , aco{j ◦ N (F ) : F ∈ ΣE })

≤ d max s(y ′ , co{j ◦ N (F ) : F ∈ ΣE }), s(y ′ , co{−j ◦ N (F ) : F ∈ ΣE })


= d max s(y ′ , {j ◦ N (F ) : F ∈ ΣE }), s(y ′ , {−j ◦ N (F ) : F ∈ ΣE }) .


28 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

In particular, if x′ ∈ BX ′ , then
s(x′ , M (E)) = hex′ , j ◦ M (E)i ≤ d s(ex′ , aco{j ◦ N (F ) : F ∈ ΣE })
= d max{s(ex′ , {j ◦ N (F ) : F ∈ ΣE }), s(ex′ , {−j ◦ N (F ) : F ∈ ΣE })}
= d max{s(ex′ , {j ◦ N (F ) : F ∈ ΣE }), s{hex′ , −j ◦ N (F )i : F ∈ ΣE })}
= d max{s(x′ , N (F )) : F ∈ ΣE }), s{hx′ , −N (F )i : F ∈ ΣE })}
= d max{s(x′ , R(NE ) ), s(x′ , R(−NE ) )} ≤ d s(x′ , R(NE ) ∪ R(−NE ) )
≤ d s(x′ , aco [R(NE ) ∪ R(−NE )] ) = s(x′ , d aco [R(NE ) ∪ R(−NE )] ).
In particular, if x ∈ M (E), then hx′ , xi ≤ s(x′ , d aco [R(NE ) ∪ R(−NE )] ).
In virtue of the Hahn-Banach theorem x ∈ d aco [ R(NE ) ∪ R(−NE )]. Thus,
M (E) ⊂ d aco [ R(NE ) ∪ R(−NE )] .
The example given in Remark 4.6 proves that in case of arbitrary mul-
timeasures the condition sub is sometimes essentially weaker than (RNj ).
The current example shows that also in case of pointless multimeasures the
sub does not guarantee the existence of the Radon-Nikodým derivative. It
is enough to take a pointless multimeasure N : Σ → cwk(X) and define
M : Σ → cwk(X) by M (E) := aco [ R(NE ) ∪ R(−NE )].
Next theorem provides a Radon-Nikodým representation without invoking
to the Rådström embedding.
Theorem 5.9. Let M, N : Σ → cb(X) be two consistent dH -multimeasures.
If M is s-usac (s-usd or s-uss) with respect to N , then there exists a non-
negative, measurable, bounded, BDS m -integrable with respect to N function
θ : Ω → R such that
Z
M (E) = θ dN, ∀ E ∈ Σ.
E

Proof. By Theorem 5.6 the s-usac (s-usd, s-uss) condition on M and N is

equivalent to property (RNj ). Then, for every x′ ∈ BX ′ , it is
Z
hex′ , j ◦ M (E)i = θ dhex′ , j ◦ N i
E

which is equivalent to
Z
′
s(x , M (E)) = θ ds(x′ , N ).
E

Then, by Proposition 2.8, θ is non-negative N -a.e.. Now, by (RNj ) and

Theorem 5.2, we have that θ is BDS m -integrable with respect to N and
Z
M (E) = θdN.
E


 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 29

Corollary 5.10. Let M, N : Σ → cb(X) satisfy the hypotheses of Theorem

5.9. If j ◦ M can be represented as a BDS-integral with respect to j ◦ N
with a non-negative bounded density θ, then also M can be represented as
the BDS m -integral of θ with respect to N .

6. Examples
Example 6.1. Let µ be a finite measure on (Ω, Σ) and N (E) := [0, µ(E)]
be an interval valued multimeasure (for results and applications of this kind
of multimeasure see for example [10, 13] and the references therein). We
can observe that if f is a scalar, bounded measurable function, then f is
BDSm-integrable with respect to N and
 Z Z 
− +
Mf (E) = − f dµ, f dµ .
E E

Example 6.2. Let let N be as in Example 6.1 and let now M (E) :=
[0, ν(E)] Zwhere ν is a finite measure on (Ω, Σ), equivalent with µ. Let
S
ν(E) = θ dµ for all E ∈ Σ. Then there exists a partition Ω = n Ωn
E
such that 0 ≤ θ ≤ n µ-a.e. on Ωn .
Moreover let ξ : R → R be defined by: ξ(a) = a if a > 0, otherwise
ξ(a) = 0. So for every a ∈ R s(a, N (E)) = ξ(a)µ(E). If α, β ∈ R and
x′ = a, y ′ = b ∈ R then
|αs(a, N ) + βs(b, N )|(E) = |αξ(a) + βξ(b)|µ(E) and
|αs(a, M ) + βs(b, M )|(E) = |αξ(a) + βξ(b)|ν(E) .
It follows that if E ∈ ΣΩn , then
ν(E)
|αs(a, M )+βs(b, M )|(E) = |αs(a, N )+βs(b, N )|(E)· ≤ n|αs(a, N )+βs(b, N )|(E) .
µ(E)
Therefore the multimeasure M is locally usd with respect to the multimea-
sure N and can be represented Zas a BDSm -integral with respect to N . One
can easily check that M (E) = θ dN .
E

Example 6.3. Assume that X is a Banach space and consider ([0, 1], L, λ)
where λ is Lebesgue measure and L is the family of all Lebesgue measurable
subsets of [0, 1]. Let f, g : [0, 1] → X be two Pettis integrable functions and
Γ(t) := co{0, f (t)}, ∆(t) := co{0, g(t)} be the multifunctions determined by
f and g respectively. They are ck(X)-valued and Pettis integrable (see [7,
Propositions 2.3 and 2.5]). We observe that, for every x′ ∈ X ′ ,
(22) s(x′ , Γ(t)) = hx′ , f i+ (t), s(x′ , ∆(t)) = hx′ , gi+ (t) .
If M, N : L → cwk(X) are the indefinite multivalued Pettis integrals of f, g,
then, ∀ x′ ∈ X ′ , ∀ E ∈ L,
Z Z
(23) s(x′ , M (E)) = hx′ , f i+ dλ, & s(x′ , N (E)) = hx′ , gi+ dλ.
E E
30 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

Assume that there exists a measurable scalar function θ which is BDSm -

integrable with respect to N and
Z
M (E) = θ dN, ∀ E ∈ L.
E
Since M, N are positive, θ is non-negative. It is a consequence of Definition
2.3 that
Z
′
s(x , M (E)) = θ ds(x′ , N ) for all x′ ∈ X ′ and E ∈ L.
E
Due to (23) we have
Z Z Z
′ + ′
hx , f i dλ = θ ds(x , N ) = θhx′ , gi+ dλ for all x′ ∈ X ′ and E ∈ L.
E E E
It follows that for each x′ ∈ X ′ we have hx′ , f i = θhx′ , gi µ-a.e.
Example 6.4. Assume that X is a Banach space and µ is a non-trivial
atomless finite measure on (Ω, Σ). Let f, f2 : Ω → X be scalarly integrable
functions and r1 , r2 : Ω → (0, ∞) be µ-integrable functions. Following [24]
or [7, Example 2.13], we define for i = 1, 2 multifunctions Γi : Ω → cb(X) by
the formulae Γi (ω) = B(fi (ω), ri (ω)), where B(x, δ) is the closed ball with
its center in x and of radius δ. Then s(x′ , Γi (ω)) = hx′ , fi (ω)i + ri (ω)kx′ k
and so each Γi is scalarly integrable. If f1 , f2 are Pettis integrable, then each
Γi is Pettis integrable in cb(X). Moreover,
Z  Z Z 
(P ) Γi dµ = B (P ) fi dµ, ri dµ i = 1, 2.
E E E
Z Z
Let M (E) := (P ) Γ1 dµ and N (E) := (P ) Γ2 dµ. One can easily check
E E Z
that M and N are dH -measures. Suppose that M (E) = θ dN, E ∈ Σ,
E
i.e.
Z Z
∀ x′ ∈ X ′ , ∀ E ∈ Σ, s(x′ , M (E)) = θ + ds(x′ , N )+ θ − ds(x′ , −N ) .
E E
That yields the equality
Z Z
′ ′ ′ ′
∀ x ∈ X , ∀ E ∈ Σ, hx , f1 i dµ + kx k r1 dµ =
ZE E
Z
θhx′ , f2 i dµ + kx′ k r2 |θ| dµ .
E E
Z 
But the sets fi dµ : E ∈ Σ , i = 1, 2 are relatively weakly compact
E
and so there exists 0 6= x′0 ∈ X ′ vanishing on these sets. It follows that
r1 = r2 |θ| µ-a.e. Appealing to [24, Lemma] we find that for each x′ ∈ X ′
one has hx′ , f1 i = θhx′ , f2 i µ-a.e. (the exceptional set depends on x′ ) i.e. f1
is scalarly equivalent to θ · f2 .
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 31

One can easily check that also the reverse implication holds true: if there
exists a measurable functionZθ such that f1 is scalarly equivalent to θf2 and
r1 = r2 |θ| µ-a.e., then M = θ dN .

A similar calculation shows that j ◦ M can be represented as a BDS-

integral with respect to j ◦ N if and only if the above θ is non-negative.
If we assume only that r1 and r2 are only positive and measurable, then we
are in the local version of the RN-Theorem 5.6.

Example 6.5. If we assume in Example 6.4 that X = Z ′ and f1 , f2 are

only Gelfand integrable then Γ1 , Γ2 are weak∗ multimeasures (see [24] for
definitions). Applying now Remark 4.7, we see that
Z Z
+
∀ z ∈ Z, ∀ E ∈ Σ, s(z, M (E)) = θ ds(z, N ) + θ − ds(z, −N )
E E

if and only there is θ such that r1 = r2 · |θ| µ-a.e. and f1 is weak∗ scalarly
equivalent to θ · f2 .

Conflict of interests: The authors declare no conflict of interest.

Author Contributions: All authors have contributed equally to this work for
writing, review and editing. All authors have read and agreed to the published
version of the manuscript.

Avaibility of matherials and data: The authors confirm that the data support-
ing the results of this study are available within the article [and/or] its supplemen-
tary material in the bibliography.

Funding: This research has been accomplished within the UMI Group TAA “Ap-
proximation Theory and Applications”; it was supported by Grant “Analisi reale,
teoria della misura ed approssimazione per la ricostruzione di immagini” (2020)
of GNAMPA – INDAM (Italy) and by Ricerca di Base 2018 dell’Università degli
Studi di Perugia - ”Metodi di Teoria dell’Approssimazione, Analisi Reale, Analisi
Nonlineare e loro Applicazioni”; Ricerca di Base 2019 dell’Università degli Studi di
Perugia - ”Integrazione, Approssimazione, Analisi Nonlineare e loro Applicazioni”;
”Metodi e processi innovativi per lo sviluppo di una banca di immagini mediche per
fini diagnostici” funded by the Fondazione Cassa di Risparmio di Perugia (FCRP),
2018; ”Metodiche di Imaging non invasivo mediante angiografia OCT sequenziale
per lo studio delle Retinopatie degenerative dell’Anziano (M.I.R.A.)”, funded by
FCRP, 2019.

Acknowledgments
This is a post-peer-review, pre-copyedit version of an article published in Jour-
nal of Convex Analysis. The final authenticated version is available online at:
https://www.heldermann.de/JCA/JCA29/JCA294/jca29062.htm
32 L. DI PIAZZA, K. MUSIAL, A. R. SAMBUCINI

References
[1] Bartle, R.G., Dunford, N., Schwartz, J., Weak compactness and vector measures,
Canad. J. Math., 7, (1955) 289–305.
[2] Beer, G., Topologies on Closed and Closed Convex Sets, Mathematics and its Appli-
cations, 268, Kluwer Academic Publishers Group, Dordrecht, (1993).
[3] Brink, H. E., Maritz, P., On the extension of set-valued set functions, Real Analysis
Exchange 26 (1), (2000), 261–276.
[4] Calabuig, J. M., Gregori, P., Sánchez Pérez, E. A., Radon-Nikodým derivatives for
vector measures belonging to Köthe function spaces, J. Math. Anal. Appl. 348 (1),
(2008), 469–479.
[5] Candeloro, D., Croitoru, A., Gavrilut A., Sambucini, A.R., A multivalued version
of the Radon-Nikodým theorem, via the single-valued Gould integral, Aust. J. Math.
Anal. Appl., 15 (2), (2018), art. 9, 1–16.
[6] Candeloro, D., Di Piazza, L., Musial, K., Sambucini, A.R., Some new results on
integration for multifunction, Ricerche di Matematica, 67 (2), (2018), 361-372, Doi:
10.1007/s11587-018-0376-x.
[7] Candeloro, D., Di Piazza, L., Musial, K., Sambucini, A.R., Multifunctions determined
by integrable functions, Internat. J. Approx. Reason., 112, (2019), 140–148, Doi:
10.1016/j.ijar.2019.06.002.
[8] Candeloro, D.; Di Piazza, L., Musial, K., Sambucini, A. R., Integration of Multifunc-
tions with Closed Convex Values in Arbitrary Banach Spaces, J. Convex Anal., 27
(4), (2020) 1233–1246.
[9] Castaing, C., Valadier, M., Convex Analysis and Measurable Multifunctions, Lecture
Notes Math. 580, Springer-Verlag, Berlin-New York (1977).
[10] Costarelli, D., Croitoru, A., Gavriluţ, A., Iosif, A., Sambucini, A.R., The Riemann-
Lebesgue integral of interval-valued multifunctions, Mathematics, 8 (12), (2020), 2250;
https://doi.org/10.3390/math8122250.
[11] Costé, A., La propriété de Radon-Nikodým en intégration multivoque, C. R. Acad.
Sci. Paris Sér. A-B 280 (1975), 22, A1515–A1518.
[12] Croitoru, A., Multivalued version of Radon-Nikodym theorem, Carphatian J. Math.,
21 (1-2), (2005), 27 - 38.
[13] A. Croitoru, A. Gavrilut, A. Iosif, A.R. Sambucini, A note on convergence results for
varying interval valued multisubmeasures, Mathematical Foundation of Computing,
4 (4), (2021), 299-310, Doi: 10.3934/mfc.2021020
[14] Fernández, A., Mayoral, F., Naranjo, F., Bartle-Dunford-Schwartz integral versus
Bochner, Pettis and Dunford integrals, J. Convex Anal., 20 (2), (2013), 339–353.
[15] Godet-Thobie, C., Some results about multimeasures and their selectors. Measure
theory, Oberwolfach 1979 (Proc. Conf., Oberwolfach, 1979), pp. 112–116, Lecture
Notes in Math., 794, Springer, Berlin, 1980.
[16] Hess, C., Set-valued integration and set-valued probability theory: an overview. Hand-
book of measure theory, Vol. I, II, 617–673, Endre Pap ed., North-Holland, Amster-
dam, 2002.
[17] Hiai, F., Radon-Nikodým theorems for set-valued measures, J. Multivariate Anal., 8
(1), (1978), 96 –118.
[18] Kandilakis D. A., On the extension of multimeasures and integration with respect to
a multimeasure, Proc. Amer. Math. Soc. 116, (1992), 85–92.
[19] Lewis, D. R., Integration with respect to vector measures, Pacific J. Math., 33, (1979),
157–165.
[20] Kluvánek I., Fourier transforms of vector valued functions and measures, Studia
Math., 37 (1970), 1–12.
 REPRESENTATIONS OF MULTIMEASURES VIA BDSm -INTEGRAL 33

[21] Kluvánek I., Knowles K., Vector measures and control systems, North-Holland Math-
ematics Studies, Vol. 20. Notes on Mathematics, No. 58, North-Holland Publishing
Co., Amsterdam-Oxford; American Elsevier Publishing Co., Inc., New York, (1976).
[22] Musial, K., A Radon-Nikodým Theorem for the Bartle-Dunford-Schwartz Integral,
Atti. Sem. Mat. Fis. Univ. Modena, 41 (1), (1993), 227–233.
[23] Musial, K., Absolute continuity of vector measures, Collect. Math. XXVII (1973),
319–321.
[24] Musial, K., Pettis integrability of multifunctions with values in arbitrary Banach
spaces, J. Convex Anal. 18 (3), (2011), 769–810.
[25] Musial, K., Approximation of Pettis Integrable Multifunctions with Values in Arbi-
trary Banach Spaces, J. Convex Anal., 20 (3), (2013), 833–870.
[26] Musial, K., Multimeasures with Values in Conjugate Banach Spaces and the Weak
Radon-Nikodým Property, J. Convex Anal., 28 (3), (2021).
[27] Panchapagesan, T. V., The Bartle-Dunford-Schwartz integral. Integration with re-
spect to a sigma-additive vector measure, Mathematics Institute of the Polish Acad-
emy of Sciences. Mathematical Monographs (New Series), 69. Birkhäuser Verlag,
Basel, (2008).
[28] Rodrı́guez, J. , On control measures of multimesures, arXiv:2109.07224v1.
[29] Sánchez Pérez, E. A., Tradacete, P., Bartle-Dunford-Schwartz integration for positive
vector measures and representation of quasi-Banach lattices, J. Nonlinear Convex
Anal., 17 (2), (2016), 387–402.
[30] Wu, W., Zhang, W., Wang, R., Set Valued Bartle Integrals, J. Math. Anal. Appl.,
225, (2001), 1–20.
[31] Zaanen, A. C., Integration, North-Holland Publ. Co. - Amsterdam 1967.

Luisa Di Piazza: Department of Mathematics, University of Palermo, Via Archirafi 34,

90123 Palermo, (Italy). Email: luisa.dipiazza@unipa.it, Orcid ID: 0000-0002-9283-5157
Kazimierz Musial : Institut of Mathematics, Wroclaw University, Pl. Grunwaldzki 2/4,
50-384 Wroclaw, (Poland). Email: kazimierz.musial@math.uni.wroc.pl, Orcid ID: 0000-
0002-6443-2043
Anna Rita Sambucini (corresponding author): Department of Mathematics and Computer
Sciences, 06123 Perugia, (Italy). Email: anna.sambucini@unipg.it, Orcid ID: 0000-0003-
0161-8729; ResearcherID: B-6116-2015.