Mediterr. J. Math.

(2023) 20:141
https://doi.org/10.1007/s00009-023-02358-6
1660-5446/23/030001-20
published online February 22, 2023
c The Author(s) 2023

Convolution Integral Operators in Variable

Bounded Variation Spaces
Laura Angeloni , Nelson J. Merentes and Maira A. Valera-López

Abstract. Working in the frame of variable bounded variation spaces

in the sense of Wiener, introduced by Castillo, Merentes, and Rafeiro,
we prove convergence in variable variation by means of the classical
convolution integral operators. In the proposed approach, a crucial step
is the convergence of the variable modulus of smoothness for absolutely
continuous functions. Several preliminary properties of the variable p(·)-
variation are also presented.
Mathematics Subject Classification. Primary 26A45; Secondary 41A35,
47G10.
Keywords. Convolution integral operators, bounded variation spaces with
variable exponent, convergence in variable variation, modulus of smooth-
ness.

1. Introduction
The study of variable exponent Lebesgue spaces has been a challenging topic
in the last 30 years. Such spaces, introduced by Orlicz [24] and then developed
by Nakano [22,23], are a generalization of the classical Lebesgue spaces: the
basic idea is that the constant exponent p of the Lp -spaces is replaced by a
variable function p(·). Such spaces share, with the classical Lp -spaces, several
properties, but nevertheless they also present some significant differences.
Among them, for example, the variable exponent spaces are not invariant
under translation. The study of such spaces had a wide development for
their intrinsic interest, and also for the important applications that they have
in partial differential equations, calculus of variations, harmonic analysis,
as well as in several applied problems such as, for example, digital image
processing (see, e.g., [11,15,28]) or the study of electrorheological fluids (see,
e.g., [25,26]).
Following the idea of variable spaces, in [14], Castillo, Merentes, and
Rafeiro introduced the variable bounded variation spaces in the sense of
Wiener (BV p(·) ), a generalization of the spaces of bounded p-variation [27],
141 Page 2 of 20 L. Angeloni et al. MJOM

that generalized in turn the classical BV-space in the sense of Jordan. We

recall that the space of bounded p-variation in the sense of Wiener is defined
as the space of functions for which the p-variation is finite, that is
n

V p [f ] := sup |f (ti ) − f (ti−1 )|p < +∞,
i=1

where the supremum is taken over all the possible increasing sequences t0 <
t1 < · · · < tn in R. Taking p = 1, the above space reduces to the classical
BV-space in sense of Jordan. The idea of Castillo, Merentes, and Rafeiro
was to replace p by a variable function p(·) with suitable properties, defining
therefore the BV p(·) -spaces, namely the variable bounded variation spaces in
the sense of Wiener, that are the setting of the present paper. We recall that
a variable exponent version of the Riesz variation was introduced and studied
in [12,13], while we refer to [7] for an extensive treatment about classical and
non-classical BV-spaces.
Our main goal will be to obtain a convergence result for the classical con-
volution integral operators with respect to convergence in variable variation
in the sense of Wiener, recalling that convergence in variation is the natural
notion of convergence in BV-spaces. Convergence results for the convolution
integral operators within BV-spaces were obtained using several notions of
variation, besides the classical Jordan variation (see, e.g., [8,9]), such as the
ϕ-variation in the sense of Musielak–Orlicz [9,21], the Riesz ϕ-variation [1],
or, in the multidimensional setting, the Tonelli variation [8]. About variable
spaces, in [16], there are results about pointwise and norm convergence for
convolution operators in the variable Lebesgue spaces Lp(·) .
As mentioned before, if variable spaces share several properties with
classical Lebesgue spaces, there are also significant differences and some im-
portant properties do not hold any more. As an example, it is not true that the
translation operator applied to a function belonging to a variable Lebesgue
space belongs to the same space, as it holds in Lp -spaces [16], and of course
the same happens in BV p(·) -spaces (see Example 3). Another delicate point
is about the additivity of the variation on intervals: analogously to what hap-
pens in the case of Musielak–Orlicz ϕ-variation (see [21, 1.17 and 1.18]), the
classical additivity property on intervals is replaced by suitable inequalities
(see Proposition 3.2). These facts make the problem of convergence in vari-
able variation much more delicate with respect to working with the classical
variation.
The paper is organized as follows. After a preliminary section in which
we state the main notations and preliminaries, we present some properties
of the variable variation in the sense of Wiener that will be useful in the
following (Sect. 3). Then, in Sect. 4, we present the main results: starting
from an estimate in variable variation for the convolution operators, we prove
a result of convergence for the modulus of smoothness, that is naturally
reformulated in the context of BV p(·) -spaces: to do this, several preliminary
results are necessary to provide a kind of approximation by means of step-
type functions (Proposition 4.3, Theorem 4.4). As a consequence, we obtain
MJOM Convolution Integral Operators Page 3 of 20 141

the convergence in variable variation by means of convolution operators. For

all these results, it is crucial the assumption of p(·)-absolute continuity on
the function, that is, the obvious reformulation of absolute continuity in the
context of variable bounded variation spaces. This is absolutely natural, since,
also in case of the classical Jordan variation, convergence in variation can be
obtained just in the subspace of BV of the absolutely continuous functions.
A similar situation occurs also working with other concepts of variation (see,
e.g., [21] for the Musielak–Orlicz ϕ-variation, [6] for the multidimensional
ϕ-variation in the sense of Tonelli) and with other kind of operators (see,
e.g., [2,10] for Mellin integral operators or [3–5] for sampling-type discrete
operators).

2. Notations and Preliminaries

We first recall the definition of variable bounded variation spaces in the sense
of Wiener, adapting to the case of functions defined on the whole real line
that one given in [14] for functions defined on an interval [a, b].
Definition 2.1. An admissible function is a function p : R → [1, +∞), such
that p+ := supx∈R p(x) < +∞.
We will also use the notation p− := inf x∈R p(x): notice that, obviously,
p− ≥ 1. From now on, p(·) will denote an admissible function.
Definition 2.2. Let f : R → R. The p(·)-variation in the Wiener’s sense of f
is defined as
n

V p(·) [f ] := sup |f (ti ) − f (ti−1 )|p(xi−1 ) ,
Π∗ i=1

where Π∗ is a tagged sequence, i.e., an increasing sequence t0 < t1 < · · · < tn ,

together with a finite sequence of numbers x0 , x1 , . . . , xn−1 subject to the
condition ti ≤ xi ≤ ti+1 , ∀i = 0, . . . , n − 1.
Definition 2.3. By
BV p(·) (R) := {f : R → R : V p(·) [λf ] < +∞, for some λ > 0},
we denote the space of functions of bounded p(·)-variation on R.
The definition in case of functions defined on an interval [a, b] ⊂ R (or
on a halfline) is given in a similar way [14].
Definition 2.4. Let f : [a, b] → R. The p(·)-variation in the Wiener’s sense of
f is defined as
n

V p(·) [f, [a, b]] := sup |f (ti ) − f (ti−1 )|p(xi−1 ) ,
Π∗ i=1

where Π∗ is a tagged partition, i.e., a partition t0 = a < t1 < · · · < tn = b of

[a, b] together with a finite sequence of numbers x0 , x1 , . . . , xn−1 subject to
the condition ti ≤ xi ≤ ti+1 , ∀i = 0, . . . , n − 1.
141 Page 4 of 20 L. Angeloni et al. MJOM

Definition 2.5. By
BV p(·) ([a, b]) := {f : [a, b] → R : V p(·) [λf, [a, b]] < +∞, for some λ > 0},
we denote the space of functions of bounded p(·)-variation in the Wiener’s
sense on [a, b].
In [14], BV p(·) ([a, b]) is actually defined by means of the norm
   
p(·) f
f BV p(·) ([a,b]) := inf λ > 0 : V , [a, b] ≤ 1 ,
λ
i.e., as the space of functions f : [a, b]→R for which f BV p(·) ([a,b]) <+∞. The
reason is that, regarding such definitions in the theory of modular spaces, it
can be proved (see [14]) that f BV p(·) ([a,b]) is a Luxemburg norm; there-
fore, BV p(·) ([a, b]) is a Banach space. Instead of the norm, we choose to use
V p(·) [λf ], that turns out to be a pseudomodular1 (see [14]), so that BV p(·) (R)
is a modular space (see, e.g., [20,21]).
The natural convergence is therefore the so-called “modular conver-
gence”.
Definition 2.6. A family of functions (fw )w>0 ⊂ BV p(·) ([a, b]) is convergent
in variation (modular convergent) to f ∈ BV p(·) ([a, b]) if there exists λ > 0,
such that
V p(·) [λ(fw − f ), [a, b]] → 0, as w → +∞.
Besides modular convergence, that is the notion of convergence that we
will use in the present paper, the norm f BV p(·) ([a,b]) induces the usual norm
convergence. We recall that norm convergence (i.e., fw − f BV p(·) ([a,b]) → 0
as w → +∞) is equivalent to
V p(·) [λ(fw − f ), [a, b]] → 0, as w → +∞, for every λ > 0.
In general, norm convergence is stronger than modular convergence: in
case of p+ < +∞, as assumed here, it can be proved that actually they are
equivalent (see also [16]).
Proposition 2.7. Given a family of functions (fw )w ⊂ BV p(·) ([a, b]), then
(fw )w converges in variation to f ∈ BV p(·) ([a, b]) if and only if (fw )w con-
verges in norm to f .
Proof. Being obvious that convergence in norm implies convergence in varia-
tion, we prove the converse. Let λ̄ > 0 be such that V p(·) [λ̄(fw −f ), [a, b]] → 0,
as w → +∞. Let us fix any λ > 0 and assume w.l.g. that λ > λ̄, since in the
1 We recall that, if X is a vector space on K (K = C or K = R), then a convex, left-
continuous function ρ : X → [0, ∞) is called a convex pseudomodular on X if, for every x
and y in X
(i) ρ(0) = 0,
(ii) ρ(αx) = ρ(x), for all α ∈ K, such that |α| = 1,
(iii) ρ(αx + (1 − α)y) ≤ αρ(x) + (1 − α)ρ(y) for all α ∈ [0, 1].
If ρ is a pseudomodular on X, then the set Xρ = {x ∈ X : limλ→0+ ρ(λx) = 0} is a
modular space.
MJOM Convolution Integral Operators Page 5 of 20 141

other case, the proof is obvious. p+ For a fixed  > 0, let w̄ > 0 be such that
V p(·) [λ̄(fw − f ), [a, b]] <  λλ̄ , for every w ≥ w̄. Then
 
p(·) p(·) λ
V [λ(fw − f ), [a, b]] = V λ̄(fw − f ), [a, b]
λ̄
p
λ + p(·)
≤ V [λ̄(fw − f ), [a, b]] < ,
λ̄
for every w ≥ w̄. 
We recall that it is easy to prove that (see [14]), if p(·) ≤ q(·), then
V p(·) [λf ] ≥ V q(·) [λf ], (2.1)
and therefore, BV p(·) (R) ⊂ BV q(·) (R).
We will study approximation properties in BV p(·) (R) of the classical
convolution integral operators defined as

(Tw f )(s) = Kw (t)f (s − t) dt, w > 0, s ∈ R,

R
for f : R → R of bounded p(·)-variation, where (Kw )w>0 is a family of kernel
functions that satisfy the usual assumptions of approximate identities, that
is
(K1) Kw ∈ L1 (R), Kw 1 ≤ A, for some constant A > 0 and for every
w > 0 and Kw (t) dt = 1, for every w > 0;
R

(K2) for any fixed δ > 0, |Kw (t)| dt → 0, as w → +∞.

|t|>δ
To get the main convergence result, it is necessary to introduce the con-
cept of variable absolute continuity, adapted from [14] to the case of functions
defined on the whole real space.
Definition 2.8. A function f ∈ BV p(·) (R) is absolutely p(·)-continuous if
n
lim sup |λ[f (ti ) − f (ti−1 )]|p(xi−1 ) = 0,
δ→0+ Πδ
i=1
for some λ > 0, where Πδ is a tagged sequence with mesh not greater than δ
(max1≤i≤n |ti − ti−1 | ≤ δ). By AC p(·) (R), we will denote the space of all the
absolutely p(·)-continuous functions f : R → R.
It is immediate to see that, in the particular case p(·) = 1, the above
definition reduces (for λ = 1) to the classical absolute continuity, expressed in
terms of convergence of the modulus of continuity, i.e., limδ→0+ ωδ1 (f ) = 0. In
general, denoted by AC(R), the space of absolutely continuous functions on R,
n
i.e., the BV -functions on R for which limδ→0+ supSδ i=1 |f (ti ) − f (ti−1 )| =
0, where the supremum is taken over all the increasing sequences Sδ on R
with mesh not greater than δ, we have that AC(R) ⊂ AC p(·) (R). Indeed,
first of all recall that p(·) ≥ 1 implies BV (R) ⊂ BV p(·) (R). Moreover, if
f ∈ AC(R), in correspondence to 0 <  < 1 (w.l.g.), there exists δ̄ > 0, such
n
that, for every 0 < δ < δ̄, supSδ i=1 |f (ti ) − f (ti−1 )| <  < 1, and so, in
141 Page 6 of 20 L. Angeloni et al. MJOM

n
particular, |f (ti ) − f (ti−1 )| ≤ i=1 |f (ti ) − f (ti−1 )| < 1, for every sequence
t0 < t1 < · · · < tn with mesh not greater than δ. Now, if Πδ is a tagged
sequence with mesh not greater than δ
|f (ti ) − f (ti−1 )|p(xi−1 ) ≤ |f (ti ) − f (ti−1 )|,
since p(·) ≥ 1, and so
n
 n

sup |f (ti ) − f (ti−1 )|p(xi−1 ) ≤ sup |f (ti ) − f (ti−1 )| < .
Πδ i=1 Sδ i=1

3. Some Properties of the p(·)-Variation

In this section, we will present some general results about the p(·)-variation
that will be useful to study the problem of convergence in variation by
means of convolution integral operators. For other basic properties of the
p(·)-variation, we refer to [17–19].

Proposition 3.1. If f ∈ BV p(·) (R), there exists λ > 0, such that

(a) V p(·) [λf ] ≤ 1;
(b) for every increasing sequence t0 < t1 < · · · < tn , λ|f (ti ) − f (ti−1 )| ≤ 1,
for every i = 1, . . . , n.
Proof. To prove (a), let μ > 0 be such that V p(·) [μf ] < +∞; if V p(·) [μf ] ≤ 1,
p(·)
there is nothing to prove. If V p(·) [μf ] > 1, then V p(·) [μf ] ≥ V p(·) [μf ],
since p(·) ≥ 1, and so, if t0 < t1 < · · · < tn , x0 , . . . , xn−1 is a tagged sequence
n  p(xi−1 ) n p(xi−1 )
μ i=1 [μ|f (ti ) − f (ti−1 )|]
|f (ti ) − f (ti−1 )| ≤ ≤ 1;
i=1
V p(·) [μf ] V p(·) [μf ]
therefore, passing to the supremum over all the tagged sequences in R, we
conclude that
V p(·) [λf ] ≤ 1
μ
for λ = V p(·) [μf ]
. About (b), it is sufficient to notice that by (a)

[λ|f (ti ) − f (ti−1 )|]p(·) ≤ V p(·) [λf ] ≤ 1,

and hence
λ|f (ti ) − f (ti−1 )| ≤ 1,
for every i = 1, . . . , n. 
The following proposition is a generalization to the p(·)-variation of the
classical additivity of the variation on intervals.

Proposition 3.2. If f ∈ BV p(·) ([a, b]) and a < c < b, then, for some λ > 0
(a) V p(·) [λf, [a, c]] + V p(·) [λf, [c, b]] ≤ V p(·) [λf, [a, b]];
2  
(b) V p(·)p+ /p− [λf, [a, b]] ≤ 2p+ /p− −1 V p(·) [λf, [a, c]] + V p(·) [λf, [c, b]] .
MJOM Convolution Integral Operators Page 7 of 20 141

Proof. Let λ > 0 be given by Proposition 3.1. For (a), it is sufficient to notice
that, if t0 = a < t1 < · · · < tm = c, x0 , . . . , xm−1 is a tagged partition of [a, c]
    
and t0 = c < t1 < · · · < tk = b, x0 , . . . , xk−1 is a tagged partition of [c, b];
 
obviously, the union t0 = a < · · · < tk = b, x0 , . . . , xk−1 is a tagged partition
of [a, b], and hence

m 
k
  
[λ|f (tj ) − f (tj−1 )|]p(xj−1 ) + [λ|f (tj ) − f (tj−1 )|]p(xj−1 ) ≤ V p(·) [λf, [a, b]].
j=1 j=1

Therefore, passing to the supremum over all the tagged partitions of [a, c]
and [c, b]
V p(·) [λf, [a, c]] + V p(·) [λf, [c, b]] ≤ V p(·) [λf, [a, b]].
To prove (b), let us consider a tagged partition of [a, b] τ0 = a < · · · <
τm = b, x0 , . . . , xm−1 . There will be some interval, say [τj−1 , τj ], that contains
c. By the convexity of the power function up(·)p+ /p− , u ≥ 0, there holds

[λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p−

= [λ|f (τj ) − f (c) + f (c) − f (τj−1 )|]p(xj−1 )p+ /p−
1 
≤ [2λ|f (τj ) − f (c)|]p(xj−1 )p+ /p− + [2λ|f (c) − f (τj−1 )|]p(xj−1 )p+ /p−
2 
2
≤ 2p+ /p− −1 [λ|f (τj ) − f (c)|]p(xj−1 )p+ /p− + [λ|f (c) − f (τj−1 )|]p(xj−1 )p+ /p− .

Now if, for example, xj−1 ∈ [τj−1 , c[, taking into account that λ|f (τj ) −
f (c)| ≤ 1 and λ|f (c) − f (τj−1 )| ≤ 1, then for some x̄c ∈ [c, τj ], we have
[λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p−
2

≤ 2p+ /p− −1 [λ|f (τj ) − f (c)|]p(x̄c ) + [λ|f (c) − f (τj−1 )|]p(xj−1 ) ,

since p(xj−1 )/p− ≥ 1 and p+ ≥ p(x̄c ). Taking into account that λ|f (τi ) −
f (τi−1 )| ≤ 1, for every i, and that p(·)p+ /p− ≥ p(·), this implies that
m

[λ|f (τi ) − f (τi−1 )|]p(xj−1 )p+ /p−
i=1

= [λ|f (τi ) − f (τi−1 )|]p(xi−1 )p+ /p− + [λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p−
i=j
2 
≤ 2p+ /p− −1 [λ|f (τi ) − f (τi−1 )|]p(xi−1 )
i=j

p2+ /p− −1
+2 [λ|f (τj ) − f (c)|]p(x̄c ) + [λ|f (c) − f (τj−1 )|]p(xj−1 )
2

≤ 2p+ /p− −1 V p(·) [λf, [a, c]] + V p(·) [λf, [c, b]] ,
and the thesis follows passing to the supremum over all the tagged partitions
of [a, b]. 
The following Proposition is a generalization of the previous result in
case of functions that vanish on a partition of [a, b].
141 Page 8 of 20 L. Angeloni et al. MJOM

Proposition 3.3. Let f ∈ BV p(·) ([a, b]) and let t0 = a < t1 < · · · < tn = b be
a partition of [a, b]. Then, for some λ > 0
n p(·)
(a) i=1 V [λf, [ti−1 , ti ]] ≤ V p(·) [λf, [a, b]];
2 n
(b) V p(·)
[λf, [a, b]] ≤ np+ /p− −1 i=1 V p(·) [λf, [ti−1 , ti ]];
(c) if, in addition, f (ti ) = 0 for every i = 0, . . . , n, V p(·)p+ /p− [λf, [a, b]] ≤
2 n
2p+ /p− −1 i=1 V p(·) [λf, [ti−1 , ti ]].

Proof. Let λ > 0 be given by Proposition 3.1. Part (a) follows with analogous
reasonings to (a) of Proposition 3.2.
To prove (b), let us consider a tagged partition of [a, b] τ0 = a < · · · <
τm = b, x0 , . . . , xm−1 . There will be some intervals [τj−1 , τj ] that contain
some ti , say τj−1 ≤ ti < · · · < ti+νj ≤ τj : for such intervals, there holds, by
the convexity of the power function up(·)p+ /p− , u ≥ 0
[λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p−
1
≤ [(νj + 1)λ|f (τj−1 ) − f (ti )|]p(xj−1 )p+ /p−
νj + 1
+ [(νj + 1)λ|f (ti ) − f (ti+1 )|]p(xj−1 )p+ /p−

+ · · · + [(νj + 1)λ|f (ti+νj ) − f (τj )|]p(xj−1 )p+ /p− (3.1)
p2+ /p− −1
≤ (νj + 1) [λ|f (τj−1 ) − f (ti )|]p(xj−1 )p+ /p−
+ [λ|f (ti ) − f (ti+1 )|]p(xj−1 )p+ /p−

+ · · · + [λ|f (ti+νj ) − f (τj )|]p(xj−1 )p+ /p− .

Now, taking into account that λ|f (τj−1 ) − f (ti )| ≤ 1, λ|f (ti ) − f (ti+1 )| ≤
1, . . . , λ|f (ti+νj ) − f (τj )| ≤ 1, then
2
[λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p− ≤ (νj + 1)p+ /p− −1 [λ|f (τj−1 ) − f (ti )|]p(x̄i−1 )

+ [λ|f (ti+1 ) − f (ti )|]p(x̄i ) + · · · + [λ|f (ti+νj ) − f (τj )|]p(x̄i+νj )

for some x̄i−1 ∈ [τj−1 , ti ], x̄k−1 ∈ [tk−1 , tk ], k = i + 1, . . . , i + νj , x̄i+νj ∈

[tij +ν , τj ], since p(xj−1 )/p− ≥ 1 for every j = 1, . . . , n, and p+ ≥ p(·). This
implies that
[λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p−
2

≤ (νj + 1)p+ /p− −1 V p(·) [λf, [τj−1 , ti ]] + · · · + V p(·) [λf, [ti+νj , τj ]] ;

hence, summing over j = 1, . . . , m, applying (a), and taking into account

that νj ≤ n − 1, for every j
m
 n

p(xj−1 )p+ /p− p2+ /p− −1
[λ|f (τj ) − f (τj−1 )|] ≤n V p(·) [λf, [ti−1 , ti ]].
j=1 i=1

Therefore, the inequality follows passing to the supremum over all the tagged
partitions of [a, b].
MJOM Convolution Integral Operators Page 9 of 20 141

To prove (c), one can proceed as in the previous case, for a tagged
partition of [a, b] τ0 = a < · · · < τm = b, x0 , . . . , xm−1 . Then, for the intervals
[τj−1 , τj ] that contain some ti , say τj−1 ≤ ti < · · · < ti+νj ≤ τj , the estimate
(3.1) can be replaced by the following:
[λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p−
1 
≤ [2λ|f (τj−1 )|]p(xj−1 )p+ /p− + [2λ|f (τj )|]p(xj−1 )p+ /p−
2 
2
≤ 2p+ /p− −1 [λ|f (τj−1 )|]p(xj−1 )p+ /p− + [λ|f (τj )|]p(xj−1 )p+ /p−
2
= 2p+ /p− −1 [λ|f (τj−1 ) − f (ti )|]p(xj−1 )p+ /p−
+ [λ|f (ti ) − f (ti+1 )|]p(xj−1 )p+ /p−

+ · · · + [λ|f (ti+νi ) − f (τj )|]p(xj−1 )p+ /p− .
Now, with similar reasonings as before, it is possible to conclude that
m
 n

2
[λ|f (τj ) − f (τj−1 )|]p(xj−1 )p+ /p− ≤ 2p+ /p− −1 V p(·) [λf, [ti−1 , ti ]],
j=1 i=1

and the inequality follows passing to the supremum over all the tagged par-
titions of [a, b].

As an immediate consequence of the previous Proposition, we have the
following:
Corollary 3.4. If f ∈ BV p(·) (R) and t0 < t1 < · · · < tn is an increasing
sequence in R, such that f (ti ) = 0 for every i = 0, . . . , n, then, for some
λ>0
n
(a) V p(·) [λf, (−∞, t0 ]]+ i=1 V p(·) [λf, [ti−1 , ti ]]+V p(·) [λf, [tn , +∞)] ≤ V p(·)
[λf ];
2 n
(b) V p(·)p+ /p− [λf ] ≤ 2p+ /p− −1 V p(·) [λf, (−∞, t0 ]]+ i=1 V p(·) [λf, [ti−1 , ti ]]

+ V p(·) [λf, [tn , +∞)] .

Another classical result for the variation that can be extended to the
p(·)-variation is the subadditivity with respect to functions:
Proposition 3.5. If f1 , . . . , fm ∈ BV p(·) (R), m ∈ N, then f1 + · · · + fm ∈
BV p(·) (R) and, for some λ > 0

V p(·) [λ(f1 + · · · + fm )] ≤ mp+ −1 V p(·) [λf1 ] + · · · + V p(·) [λfm ] .

Proof. Let λ > 0 be such that V p(·) [λfi ] < +∞, for every i = 1, . . . , m, and
let t0 < t1 < · · · < tn , x0 < · · · < xn−1 be a tagged sequence. Then, by the
monotonicity and the convexity of the power function up(·) , u ≥ 0
n
|λ(f1 + · · · + fm )(ti ) − λ(f1 + · · · + fm )(ti−1 )|p(xi−1 )
i=1
141 Page 10 of 20 L. Angeloni et al. MJOM

n
 p(xi−1 )
≤ [λ|f1 (ti ) − f1 (ti−1 )| + · · · + λ|fm (ti ) − fm (ti−1 )|]
i=1
n 
 1 p(xi−1 )
≤ mλ|f1 (ti ) − f1 (ti−1 )| + ···
i=1
m
1 p(xi−1 ) 
+ mλ|fm (ti ) − fm (ti−1 )|
m

≤ mp+ −1 V p(·) [λf1 ] + · · · + V p(·) [λfm ] ,

and the thesis follows passing to the supremum over all the tagged sequences
in R. 

Finally, we prove a relation between the p(·)-variations of a function

and its shifted version τt f (u) := f (u − t), t, u ∈ R, that will be fundamental
to work with the convolution integral operators.

Proposition 3.6. For every t ∈ R

V p(·) [τt f ] = V τ−t p(·) [f ].
Therefore, τt f ∈ BV p(·) (R), t ∈ R, if and only if f ∈ BV τ−t p(·) (R).

Proof. Let s0 < s1 < · · · < sn , x0 < · · · < xn−1 be a fixed tagged sequence:
then, for t ∈ R, s0 − t < s1 − t < · · · < sn − t, x0 − t < · · · < xn−1 − t is again
a tagged sequence. Therefore
n
 n

p(xi−1 )
|f (si − t) − f (si−1 − t)| = |f (si − t) − f (si−1 − t)|p((xi−1 −t)+t)
i=1 i=1
p(·+t)
≤V [f ],
and so, passing to the supremum over all the tagged sequences in R
V p(·) [τt f ] ≤ V τ−t p(·) [f ].
On the other side, if s0 < s1 < · · · < sn , x0 < · · · < xn−1 is a tagged
sequence, so is s0 + t < s1 + t < · · · < sn + t, x0 + t < · · · < xn−1 + t, and
therefore
n
 n

|f (si ) − f (si−1 )|p(xi−1 +t) = |f ((si + t) − t) − f ((si−1 + t) − t)|p(xi−1 +t)
i=1 i=1

≤ V p(·) [τt f ].

Again passing to the supremum over all the tagged sequences, we have
V τ−t p(·) [f ] ≤ V p(·) [τt f ],
and the result is proved. 

Example. The previous Proposition suggests an important difference between

the variable variation and the classical notion of variation: the space BV p(·) (R)
MJOM Convolution Integral Operators Page 11 of 20 141

is not invariant under translation, as the classical BV-spaces. Indeed, let us

consider for example the function f : R → R defined as

x, x = k1 , k = 1, 2, . . . ,
f (x) =
0, otherwise,
and the admissible function

2, x ∈ [0, 1],
p(x) =
1, otherwise.
Then, f ∈ BV p(·) (R), since
+∞ 
 2 +∞

 1  1
V p(·) [f ] = 2 f = 2 < +∞.
 k  k2
k=1 k=1

Nevertheless, if we consider h ∈ R, such that |h| > 1, then by Proposition

3.6, for every λ > 0
n
 +∞
 λ
V p(·) [λτh f ] = V p(·−h) [λf ] = sup |λf (ti ) − λf (ti−1 )|p(xi −h) = 2 = +∞,
Π∗ i=1 k
k=1

taking into account that, for every xi ∈ [0, 1], xi − h ∈ [0, 1]. Therefore,
τh f ∈ BV p(·) (R).

4. Convergence Results by Means of Convolution Operators

We first point out that, if f ∈ BV p(·) (R), the operators Tw f are well defined.
Indeed, it is immediate to see that f ∈ BV p(·) (R) implies that f is bounded,
and therefore, for some M > 0

|(Tw f )(s)| ≤ |Kw (t)||f (s − t)| dt ≤ M Kw 1 ≤ M A,

R
for every s ∈ R, w > 0, by (K1).
The first result will be an estimate in variable variation for Tw f , w > 0.
Theorem 4.1. If f ∈ BV p(·) (R) and (K1) is satisfied, then there exists μ > 0,
such that
V p+ /p− p(·) [μTw f ] ≤ V p(·) [λf ], (4.1)
p(·)
where λ > 0 is such that V [λf ] < +∞. As a consequence, Tw maps
BV p(·) (R) in BV p+ /p− p(·) (R), for every w > 0.
Proof. Let s0 < s1 < · · · < sn , x0 < · · · < xn−1 be a tagged sequence in R.
Then, for μ > 0, we have
n
Sw := |μ[(Tw f )(si ) − (Tw f )(si−1 )]|p+ /p− p(xi−1 )
i=1
n 

p+ /p− p(xi−1 )
 
= μ Kw (t)f (si − t) dt − μ Kw (t)f (si−1 − t) dt .

i=1 R R
141 Page 12 of 20 L. Angeloni et al. MJOM

Now, by the convexity of the function up+ /p− p(·) , u ≥ 0, by (K1) and Jensen’s
inequality
n
Sw ≤ A−1 |Kw (t)| [μA|f (si − t) − f (si−1 − t)|]p+ /p− p(xi−1 ) dt
R i=1

−1
≤A |Kw (t)|V p+ /p− p(·) [μAτt f ] dt,
R
and so, by Proposition 3.6

Sw ≤ A−1 |Kw (t)|V p+ /p− p(·+t) [μAf ] dt.

R
Since, for every t ∈ R, p+ /p− p(· + t) ≥ p(·)/p− p(· + t) ≥ p(·), then by (2.1)
and (K1)

Sw ≤ A−1 |Kw (t)|V p(·) [μAf ] dt ≤ V p(·) [μAf ].

R
λ
Therefore, if 0 < μ < A , passing to the supremum over all the possible tagged
sequences in R, we conclude that
V p+ /p− p(·) [μTw f ] ≤ V p(·) [λf ].

Remark 4.2. In the particular case of p(·) = 1, the variable bounded variation
reduces to the classical Jordan variation and the estimate of Theorem 4.1
becomes
V [Tw f ] ≤ V [f ],
that is, the variation-diminishing property for the classical convolution in-
tegral operators (see, e.g., [8]). In the case p(·) ≡ p, p > 1, the variable
bounded variation coincides with the Wiener p-variation and the estimate of
Theorem 4.1 reduces to the variation-diminishing property for the Wiener
p-variation (see, e.g., [21] for the generalization of such result in the case of
the Musielak–Orlicz ϕ-variation). We point out that, in all these cases, with
”classical” notions of variation, the convolution integral operators map the
space of functions of bounded variation (in the sense of Jordan, Wiener...) in
itself, while here they map BV p(·) (R) into BV p+ /p− p(·) (R). Actually, this is
natural, in the setting of variable exponent spaces: for instance, similar phe-
nomena occur working in variable Lebesgue spaces, that are not translation
invariant (see, e.g., [16]).
To prove the main convergence theorem, a crucial step is a result about
the convergence of the modulus of smoothness of an absolutely continuous
function: the modulus of smoothness, in the context of BV p(·) -spaces, for
f : R → R, is defined as
ω p(·) (f, δ) := sup V p(·) [τt f − f ], δ > 0.
|t|≤δ

We will now prove some preliminary results. The first proposition guarantees
the possibility to approximate a BV p(·) -function by means of a step function.
MJOM Convolution Integral Operators Page 13 of 20 141

Proposition 4.3. If f ∈ AC p(·) (R), there exists λ > 0, such that, for every
 > 0, there exist a, b ∈ R and δ > 0, such that, if t0 = a < t1 < · · · < tn = b
is a partition of [a, b] with ti − ti−1 < δ for every i = 1, . . . , n, then
(a) V p(·) [λf, (−∞, a]] <  and V p(·) [λf, [b, +∞)] < ;
n p(·)
(b) i=1 V [λf, [ti−1 , ti ]] < ;
(c) the step functions ν1 , ν2 : R → R defined as
⎧
⎪
⎨f (a), t < a,
ν1 (t) := f (ti−1 ), ti−1 ≤ t < ti ,
⎪
⎩
f (b), t ≥ b,
⎧
⎪
⎨f (a), t ≤ a,
ν2 (t) := f (ti ), ti−1 < t ≤ ti ,
⎪
⎩
f (b), t > b,

are such that V p(·)p+ /p− [λ(f − ν1 )] <  and V p(·)p+ /p− [λ(f − ν2 )] < .

Proof. About (a), it is sufficient to recall that f ∈ AC p(·) (R) implies in partic-
ular f ∈ BV p(·) (R), and hence, there exists λ > 0, such that V p(·) [λf ] < +∞.
Since V p(·) [λf ] = limn→+∞ V p(·) [λf, [−xn , xn ]] where (xn )n is an increasing
sequence in R, then by Corollary 3.4

V p(·) [λf, (−∞, −xn ]] + V p(·) [λf, [xn , +∞)] ≤ V p(·) [λf ] − V p(·) [λf, [−xn , xn ]] → 0,

as n → +∞. Therefore
lim V p(·) [λf, (−∞, −xn ]] = lim V p(·) [λf, [xn , +∞)] = 0,
n→+∞ n→+∞

that implies (a).

To prove (b), it is sufficient to notice that, by the p(·)-absolute continuity
of f , for some λ > 0, in correspondence to  > 0, there exists δ > 0, such
n
that i=1 |λ[f (ti ) − f (ti−1 )]|p(xi−1 ) < , for every tagged sequence t0 < t1 <
· · · < tn , such that ti − ti−1 < δ, for every i. Therefore, if one considers a
tagged sequence τ0i < · · · < τm i
i
, y0i , . . . , ym
i
i −1
in each interval [ti−1 , ti ], there
holds
mi
n 
 i
[λ[f (τji ) − f (τj−1
i
)]|p(yj−1 ) < ,
i=1 j=1

and the thesis follows passing to the supremum over all the possible tagged
sequences in [ti−1 , ti ].
Let us now prove (c). In correspondence to  > 0, let λ > 0, a, b ∈ R
and δ > 0 be given by (a) and (b), so that, if t0 = a < t1 < · · · < tn = b is a
partition of [a, b] with ti − ti−1 < δ for every i = 1, . . . , n, then
 
(i) V p(·) [λf, (−∞, a]] < p2
(< ) and V p(·) [λf, [b, +∞)] < p2
(<
2 + /p− +1 2 + /p− +1

);
n 
(ii) i=1 V p(·) [λf, [ti−1 , ti ]] < p2
(< ).
2 + /p− +p+
141 Page 14 of 20 L. Angeloni et al. MJOM

By Corollary 3.4 (taking into account that (f − νk )(ti ) = 0 for every i =

0, . . . , n, k = 1, 2)
2
V p(·)p+ /p− [λ(f − νk )] ≤ 2p+ /p− −1 V p(·) [λ(f − νk ), (−∞, a]]
n

+ V p(·) [λ(f − νk ), [ti−1 , ti ]]
i=1

+ V p(·) [λ(f − νk ), [b, +∞)]
2
= 2p+ /p− −1 V p(·) [λf, (−∞, a]]
n

+ V p(·) [λ(f − νk ), [ti−1 , ti ]]
i=1

+ V p(·) [λf, [b, +∞)]
 n

 
p2+ /p− −1 p(·)
<2 2 + V [λ(f − νk ), [ti−1 , ti ]] .
2p+ /p− i=1

By Proposition 3.5, and since obviously V p(·) [λνk , [ti−1 , ti ]] ≤ V p(·) [λf,
[ti−1 , ti ]], we have that
V p(·) [λ(f − νk ), [ti−1 , ti ]] ≤ 2p+ −1 {V p(·) [λf, [ti−1 , ti ]] + V p(·) [λνk , [ti−1 , ti ]]}
≤ 2p+ V p(·) [λf, [ti−1 , ti ]],
for every i = 1, . . . , n. Therefore, by (ii)
 
2  
V p(·)p+ /p− [λ(f − νk )] < 2p+ /p− −1 2 + 2p+ 2 = ,
2p+ /p− 2p+ /p− +p+
k = 1, 2. 
p(·)
Theorem 4.4. If f ∈ AC (R), then there exists λ > 0, such that
lim V p(·)p+ /p− [λ(τt ν1 − ν1 )] = 0, lim V p(·)p+ /p− [λ(τt ν2 − ν2 )] = 0,
t→0− t→0+
where ν1 and ν2 are defined as in Proposition 4.3.
Proof. Let us fix  > 0; by Proposition 4.3, there exist λ, δ > 0, a, b ∈ R and
two step functions ν1 : R → R, ν2 : R → R, such that
 n

V p(·) [λf, [ti−1 , ti ]] < p2 /p +p −1 , (4.2)
− +
i=1 2 +

where t0 = a < t1 < · · · < tn = b is a partition of [a, b], such that ti −ti−1 < δ,
for every i = 1, . . . , n, and V p(·)p+ /p− [λ(f − νk )] < , k = 1, 2.
Let now 0 < β < mini=1,...,n {ti − ti−1 }. If −β < t < 0, then ti−1 <
ti−1 − t < ti and τt ν1 (ti−1 ) = ν1 (ti−1 − t) = ν1 (ti−1 ) = f (ti−1 ): therefore,
V p(·) [λτt ν1 , [ti−1 , ti ]] = V p(·) [λν1 , [ti−1 , ti ]]. Then, using (c) of Proposition
3.3, Proposition 3.5, and Eq. (4.2), we obtain
n
2
V p(·)p+ /p− [λ(τt ν1 − ν1 ), [a, b]] ≤ 2p+ /p− −1 V p(·) [λ(τt ν1 − ν1 ), [ti−1 , ti ]]
i=1
MJOM Convolution Integral Operators Page 15 of 20 141

n

2
≤ 2p+ /p− +p+ −2 V p(·) [λτt ν1 , [ti−1 , ti ]]
i=1

p(·)
+V [λν1 , [ti−1 , ti ]]
n

2
= 2p+ /p− +p+ −1 V p(·) [λν1 , [ti−1 , ti ]]
i=1
n

2
≤ 2p+ /p− +p+ −1 V p(·) [λf, [ti−1 , ti ]] < .
i=1

Moreover, taking into account that ν1 (u) = τt ν1 (u) = f (a), for every u ≤ a,
and ν(u) = τt ν(u) = f (b), for every u ≥ b
V p(·)p+ /p− [λ(τt ν1 − ν1 )] = V p(·)p+ /p− [λ(τt ν1 − ν1 ), [a, b]] < .
The proof of the other limit relation for ν2 follows with analogous
reasonings. 

We will now prove a result of convergence for the modulus of smoothness

in case of AC p(·) -functions.

Theorem 4.5. If f ∈ AC p(·) (R), then for some λ > 0, there holds
2 2
lim ω p(·)p+ /p− (λf, δ).
δ→0+

Proof. We will prove that

2 2
lim V p(·)p+ /p− [λ(τt f − f )] = 0,
t→0

that is equivalent to the thesis. For a fixed  > 0, by Proposition 4.3, there
exist λ, δ > 0, a, b ∈ R and two step functions ν1 : R → R and ν2 : R → R
(associated with a partition of [a, b] with mesh not greater than δ), such that

V p(·)p+ /p− [λ(f − νk )] < p+ , k = 1, 2. (4.3)
3
If t < 0, by Proposition 3.5
2 2 2 2
V p(·)p+ /p− [λ(τt f − f )] ≤ 3p+ −1 V p(·)p+ /p− [λ(τt f − τt ν1 )]
2 2 2 2

+V p(·)p+ /p− [λ(τt ν1 − ν1 )] + V p(·)p+ /p− [λ(ν1 − f )] .

By Eqs. (2.1) and (4.3), there holds

2 2 
V p(·)p+ /p− [λ(ν1 − f )] ≤ V p(·)p+ /p− [λ(ν1 − f )] < ,
3p+
and by Theorem 4.4
2 2 
V p(·)p+ /p− [λ(τt ν1 − ν1 )] ≤ V p(·)p+ /p− [λ(τt ν1 − ν1 )] <
,
3p+
for sufficiently small t < 0. Finally, by Proposition 3.6 and (2.1)
2 2 2 2
V p(·)p+ /p− [λ(τt f − τt ν1 )] = V p(·+t)p+ /p− [λ(f − ν1 )] ≤ V p(·)p+ /p− [λ(f − ν1 )].
141 Page 16 of 20 L. Angeloni et al. MJOM

Therefore, we conclude that for t < 0 small enough

2 2 3
V p(·)p+ /p− [λ(τt f − f )] ≤ 3p+ −1 = .
3p+
Now, replacing ν1 by ν2 , it is possible to prove that, for t > 0 sufficiently
small
2 2
V p(·)p+ /p− [λ(τt f − f )] ≤ ,
and the proof is complete. 

We are now ready to prove the main result about convergence in p(·)-
variation.

Theorem 4.6. Let f ∈ AC p(·) (R). If (K1) and (K2) are satisfied, then
2 2
lim V p+ /p− p(·) [λ(Tw f − f )] = 0,
w→+∞

for some λ > 0.

Proof. For a fixed tagged sequence s0 < s1 < · · · < sn , x0 < · · · < xn−1 and
λ > 0, there holds
n
 2 2
Sw := |λ[(Tw f − f )(si ) − (Tw f − f )(si−1 )]|p+ /p− p(xi−1 )
i=1
n 
 
= λ Kw (t)[f (si − t) − f (si )] dt

i=1 R
p2+ /p2− p(xi−1 )

−λ Kw (t)[f (si−1 − t) − f (si−1 )] dt
R
n 

≤ λ |Kw (t)||[f (si − t) − f (si )]
i=1 R
p2+ /p2− p(xi−1 )
− [f (si−1 − t) − f (si−1 )]| dt ,

taking into account of (K1). Now, by Jensen’s inequality and (K1)

n 
 
Sw ≤ A−1 |Kw (t)| λA [f (si − t) − f (si )]
R i=1
p2+ /p2− p(xi−1 )

− [f (si−1 − t) − f (si−1 )]  dt
  n 
 
= A−1 + |Kw (t)| λA [f (si − t) − f (si )]
|t|≤δ |t|>δ i=1
p2+ /p2− p(xi−1 )

− [f (si−1 − t) − f (si−1 )]  dt
:= I1δ + I2δ ,
MJOM Convolution Integral Operators Page 17 of 20 141

for every δ > 0. Let us fix  > 0. By Theorem 4.5 there exist λ̄, δ̄ > 0 such
2 2
that ω p+ /p− p(·) (λ̄f, δ̄) < 2 , and so, in correspondence of δ̄, for 0 < λ < λ̄A−1
2 2
I1δ̄ ≤ A−1 |Kw (t)|V p+ /p− p(·) [λA(τt f − f )] dt
|t|≤δ̄

p2+ /p2− p(·) 

≤ω (λAf, δ̄)A−1 |Kw (t)| dt < ,
|t|≤δ̄ 2
by (K1).
About I2δ , by Proposition 3.5 and Eq. (2.1)
2 2 2 2

I2δ ≤ A−1 2p+ −1 |Kw (t)| V p+ /p− p(·) [λAτt f ] + V p+ /p− p(·) [λAf ] dt
|t|>δ

−1 p+ p+ /p− p(·)
≤A 2 V [λAf ] |Kw (t)| dt
|t|>δ

≤ A−1 2p+ V p(·) [μf ] |Kw (t)| dt,

|t|>δ

for 0 < λ < μA−1 , where μ is such that V p(·) [μf ] < +∞.2 By assumption
(K2), there exists w̃ > 0 such that |t|>δ̄ |Kw (t)| dt < A−1 2p+ +1 V p(·) [μf ] , for
every w ≥ w̃, and so

I2δ̄ < .
2
Therefore, if 0 < λ < min{λ̄A−1 , μA−1 }
Sw ≤ I1δ̄ + I2δ̄ < ,
for sufficiently large w > 0, and the thesis follows passing to the supremum
over all the possible tagged sequences in R. 

Funding Open access funding provided by Università degli Studi di Perugia

within the CRUI-CARE Agreement.

Data Availability The manuscript has no associated data.

Declarations

Conflict of Interest The first author is member of the Gruppo Nazionale per
l’Analisi Matematica, la Probabilità e le loro Applicazioni (GNAMPA) of the
Istituto Nazionale di Alta Matematica (INdAM), of RITA (Research ITalian
network on Approximation) and of the UMI group “Teoria dell’Approssimazi-
one e Applicazioni” and is partially supported by the “Department of Math-
ematics and Computer Science” of the University of Perugia, by the “Fondo
Ricerca di Base” 2019 and 2020 of the University of Perugia, by 2018 (B.I.M.)
and 2019 (M.I.R.A.) Projects funded by the Fondazione Cassa di Risparmio

2 W.l.g. V p(·) [μf ] = 0, since otherwise the result is trivial.

141 Page 18 of 20 L. Angeloni et al. MJOM

di Perugia and by a 2022 GNAMPA-INdAM Project (“Enhancement e seg-

mentazione di immagini mediante operatori di tipo campionamento e metodi
variazionali per lo studio di applicazioni biomediche”).

Open Access This article is licensed under a Creative Commons Attribution 4.0
International License, which permits use, sharing, adaptation, distribution and re-
production in any medium or format, as long as you give appropriate credit to
the original author(s) and the source, provide a link to the Creative Commons
licence, and indicate if changes were made. The images or other third party ma-
terial in this article are included in the article’s Creative Commons licence, unless
indicated otherwise in a credit line to the material. If material is not included in
the article’s Creative Commons licence and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permis-
sion directly from the copyright holder. To view a copy of this licence, visit http://
creativecommons.org/licenses/by/4.0/.
Publisher’s Note Springer Nature remains neutral with regard to jurisdic-
tional claims in published maps and institutional affiliations.

References
[1] Angeloni, L.: A new concept of multidimensional variation in the sense of Riesz
and applications to integral operators. Mediterr. J. Math. 14(4), 149 (2017)
[2] Angeloni, L., Vinti, G.: Variation and approximation in multidimensional set-
ting for Mellin integral operators, New perspectives on approximation and sam-
pling theory, pp. 299–317, Appl. Numer. Harmon. Anal. Birkhäuser/Springer,
Cham (2014)
[3] Angeloni, L., Costarelli, D., Vinti, G.: A characterization of the convergence
in variation for the generalized sampling series. Ann. Acad. Sci. Fenn. Math.
43, 755–767 (2018)
[4] Angeloni, L., Costarelli, D., Vinti, G.: A characterization of the absolute con-
tinuity in terms of convergence in variation for the sampling Kantorovich op-
erators. Mediterr. J. Math. 16, 44 (2019)
[5] Angeloni, L., Costarelli, D., Vinti, G.: Convergence in variation for the multidi-
mensional generalized sampling series and applications to smoothing for digital
image processing. Ann. Acad. Sci. Fenn. Math. 45, 751–770 (2020)
[6] Angeloni, L., Vinti, G.: Convergence and rate of approximation for linear in-
tegral operators in BV ϕ −spaces in multidimensional setting. J. Math. Anal.
Appl. 349, 317–334 (2009)
[7] Appell, J., Banaś, J., Merentes, N.: Bounded Variation and Around, vol. 17.
De Gruyter Series in Nonlinear Analysis and Applications. De Gruyter, Berlin
(2014)
[8] Bardaro, C., Butzer, P.L., Stens, R.L., Vinti, G.: Convergence in variation and
rates of approximation for Bernstein-type polynomials and singular convolution
integrals. Analysis 23, 299–340 (2003)
[9] Bardaro, C., Musielak, J., Vinti, G.: Nonlinear Integral Operators and Appli-
cations. De Gruyter Series in Nonlinear Analysis and Applications. Walter De
Gruyter, New York (2003)
MJOM Convolution Integral Operators Page 19 of 20 141

[10] Bardaro, C., Sciamannini, S., Vinti, G.: Convergence in BVϕ by nonlinear
Mellin- type convolution operators. Funct. Approx. Comment. Math. 29, 17–
28 (2001)
[11] Blomgren, P., Chan, T., Mulet, P., Wong, C.K.: Total variation image restora-
tion: numerical methods and extensions. In: Proceedings of the 1997 IEEE
International Conference on Image Processing, vol. III, pp. 384–387 (1997)
[12] Castillo, R.E., Guzmán, O.M., Rafeiro, H.: Variable exponent bounded varia-
tion spaces in the Riesz sense. Nonlinear Anal. Theory Meth. Appl. 132, 173–
182 (2016)
[13] Castillo, R.E., Guzmán, O.M., Rafeiro, H.: Nemytskii Operator in Riesz-
Bounded Variation Spaces with Variable Exponent. Mediterr. J. Math. 14(2)
(2017)
[14] Castillo, R.E., Merentes, N., Rafeiro, H.: Bounded variation spaces with p-
variable. Mediterr. J. Math. 11(4), 1069–1079 (2014)
[15] Chen, Y., Levine, S., Rao, M.: Variable exponent, linear growth functionals in
image restoration. SIAM J. Appl. Math. 66(4), 1383–1406 (2006)
[16] Cruz-Uribe, D.V., Fiorenza, A.: Variable Lebesgue Spaces. Foundations and
Harmonic Analysis. Birkhäuser, Basel (2013)
[17] Mejı́a, O., Merentes, N., Sánchez, J.L., Valera-López, M.: The space of bounded
p(·)-variation in the sense Wiener–Korenblum with variable exponent. Adv.
Pure Math. 6, 21–40 (2016)
[18] Mejı́a, O., Merentes, N., Sánchez, J.L.: The space of bounded p(·)-variation in
Wiener’s sense with variable exponent. Adv. Pure Math. 5, 703–716 (2015)
[19] Mejı́a, O., Silvestre, P., Valera-López, M.: Functions of bounded (p(·, 2))-
variation in De La Vallée Poussin–Wiener’s sense with variable exponent. Adv.
Pure Math. 7, 507–532 (2017)
[20] Musielak, J.: Orlicz Spaces and Modular Spaces, Lecture Notes in Math, vol.
1034. Springer, Berlin (1983)
[21] Musielak, J., Orlicz, W.: On generalized variations. I. Stud. Math. 18, 11–41
(1959)
[22] Nakano, H.: Modulared Semi-Ordered Linear Spaces. Maruzen Co., Ltd., Tokyo
(1950)
[23] Nakano, H.: Topology and Linear Topological Spaces. Maruzen Co., Ltd.,
Tokyo (1951)
[24] Orlicz, W.: Über konjugierte Exponentenfolgen. Stud. Math. 3, 200–211 (1931)
[25] Ruzicka, M.: Electrorheological Fluids: Modeling and Mathematical Theory.
Lecture Notes in Mathematics, vol. 1748. Springer, Berlin (2000)
[26] Ruzicka, M.: Modeling, mathematical and numerical analysis of electrorheo-
logical fluids. Appl. Math. 49(6), 565–609 (2004)
[27] Wiener, N.: The quadratic variation of a function and its Fourier coefficients.
J. Math. Phys. 3(2), 72–94 (1924)
[28] Wunderli, T.: On time flows of minimizers of general convex functionals of
linear growth with variable exponent in BV space and stability of pseudo-
solutions. J. Math. Anal. Appl. 364(2), 5915–5998 (2010)
141 Page 20 of 20 L. Angeloni et al. MJOM

Laura Angeloni
Dipartimento di Matematica e Informatica
Università degli Studi di Perugia
Via Vanvitelli 1
06123 Perugia
Italy
e-mail: [email protected]

Nelson J. Merentes and Maira A. Valera-López

Escuela de Matemática, Facultad de Ciencias
Universidad Central de Venezuela
Caracas 1010
Venezuela
e-mail: [email protected]

Maira A. Valera-López
e-mail: [email protected]

Received: March 28, 2022.

Accepted: January 26, 2023.