Professional Documents
Culture Documents
Fractalfract 07 00579
Fractalfract 07 00579
Fractalfract 07 00579
Article
New Estimates on Hermite–Hadamard Type Inequalities via
Generalized Tempered Fractional Integrals for Convex
Functions with Applications
Artion Kashuri 1 , Yahya Almalki 2 , Ali M. Mahnashi 3 and Soubhagya Kumar Sahoo 4, *
1 Department of Mathematics, Faculty of Technical and Natural Sciences, University “Ismail Qemali”,
9400 Vlora, Albania; artion.kashuri@univlora.edu.al
2 Department of Mathematics, College of Science, King Khalid University, Abha 61413, Saudi Arabia
3 Department of Mathematics, College of Science, Jazan University, Jazan 45142, Saudi Arabia
4 Department of Mathematics, C.V. Raman Global University, Bhubaneswar 752054, India
* Correspondence: soubhagyalulu@gmail.com
Abstract: This paper presents a novel approach by introducing a set of operators known as the left
and right generalized tempered fractional integral operators. These operators are utilized to establish
new Hermite–Hadamard inequalities for convex functions as well as the multiplication of two convex
functions. Additionally, this paper gives two useful identities involving the generalized tempered
fractional integral operator for differentiable functions. By leveraging these identities, our results
consist of integral inequalities of the Hermite–Hadamard type, which are specifically designed to
accommodate convex functions. Furthermore, this study encompasses the identification of several
special cases and the recovery of specific known results through comprehensive research. Lastly,
this paper offers a range of applications in areas such as matrices, modified Bessel functions and
q-digamma functions.
Theorem 1 (H–H inequality). For a convex function S : I → R and two points µ1 , µ2 ∈ I with
µ1 < µ2 , the following double inequality is valid:
S(µ1 ) + S(µ2 )
µ1 + µ2 1
Z µ2
S ≤ S(x)dx ≤ . (2)
2 µ 2 − µ 1 µ1 2
In recent decades, researchers have investigated inequality (2) using newly formulated
definitions motivated by convex functions. For further exploration, interested readers can
refer to the following sources: [7–22].
Definition 2. For any real number α > 0 and x, ζ ≥ 0, the ζ-incomplete gamma function is
defined as Z x
γζ (α, x) := uα−1 e−ζu du.
0
In the following sections, we will revisit the definition of tempered fractional integral operators.
Definition 3 ([22]). Let S ∈ L[µ1 , µ2 ] (the set of all Lebesgue integrable functions on [µ1 , µ2 ]),
where 0 ≤ µ1 < µ2 . Then for ζ ≥ 0, the tempered fractional integral operators µ+ Iα,ζ S and µ− Iα,ζ S
1 2
of order α > 0 are defined as
Z x
1
µ1+ I
α,ζ
S(x) := (x − σ)α−1 e−ζ (x−σ) S(σ)dσ, x > µ1 (3)
Γ(α) µ1
Fractal Fract. 2023, 7, 579 3 of 19
and
1
Z µ2
µ2− I
α,ζ
S(x) := (σ − x)α−1 e−ζ (σ−x) S(σ)dσ, x < µ2 , (4)
Γ(α) x
respectively.
The function Υ : [0, +∞) → [0, +∞) was introduced by Sarikaya et al. in [18], charac-
terized by the following conditions:
Υ (d)
Z 1
dd < +∞, (7)
0 d
1 Υ(d1 ) 1 d
≤ ≤ A1 for ≤ 1 ≤ 2, (8)
A1 Υ(d2 ) 2 d2
Υ(d2 ) Υ (d )
≤ A2 21 for d1 ≤ d2 , (9)
d22 d1
Υ(d2 ) Υ(d1 ) Υ (d ) 1 d
2
− 2
≤ A3 |d2 − d1 | 22 for ≤ 1 ≤ 2, (10)
d2 d1 d2 2 d2
Υ ( d2 )
that ζ is a decreasing function, then it can be deduced from [23] that Υ satisfies (7)–(10).
d2
Based on this, the definitions of the left-sided and right-sided generalized integral operators
are given as follows:
Υ (x − σ )
Z x
µ1+ IΥ S(x) := S(σ)dσ, x > µ1 , (11)
µ1 x−σ
Υ ( σ − x)
Z µ2
µ2− IΥ S(x) :=
σ−x
S(σ)dσ, x < µ2 . (12)
x
A notable feature of generalized integrals is their ability to yield Riemann–Liouville
fractional integrals, Katugampola fractional integrals, and other types as well. These results
can be found in references such as [18,20,23,24].
In recent years, there has been a growing interest in extending the H–H inequality
to the context of fractional calculus. Fractional calculus deals with the generalization
of derivatives and integrals to non-integer orders, and it has emerged as a powerful
mathematical tool for modeling and analyzing complex phenomena in physics, engineering,
and other fields. The fractional H–H inequality explores the behavior of fractional integrals
of convex functions. By considering fractional operators of different orders, such as the
Riemann–Liouville or Caputo fractional operators, the inequality provides a bridge between
the concepts of convexity and fractional calculus. It establishes relationships between
the fractional integral of a convex function and its endpoint values, revealing intriguing
properties and insights into the behavior of these functions. The study of fractional H–H
inequalities has attracted significant attention from researchers in recent years. Many
mathematicians have worked on deriving new inequalities, investigating their properties,
Fractal Fract. 2023, 7, 579 4 of 19
and exploring their applications in various areas [25,26]. These developments have led to a
deeper understanding of the interplay between convex functions and fractional calculus.
In Section 2, we begin by introducing the left and right generalized tempered fractional
integral operators as novel tools in the study of H–H inequalities. By utilizing these
operators, we establish novel inequalities for convex functions as well as for products of
two convex functions. These results extend the existing literature and provide valuable
insights into the behavior of convex functions.
Moving on to Section 3, we delve into the exploration of two interesting identities
involving the generalized tempered fractional integral operators for differentiable functions.
Through the utilization of these identities, we obtain supplementary integral inequalities
of H–H and midpoint types for convex functions. This further expands the scope and
applicability of our findings.
Section 4 takes a practical turn as we delve into the application of our results. We
discuss their relevance and implications in various domains, such as matrices, modified
Bessel functions and q-digamma functions.
Finally, in Section 5, we wrap up the paper with a concise conclusion summarizing the
key findings and their implications. We also provide future recommendations, suggesting
potential directions for further research.
Through this structured organization, our paper presents a comprehensive analysis
of the introduced operators, establishes new inequalities, explores intriguing identities,
demonstrates practical applications, and offers valuable insights for future research.
Definition 4. The operator that defines the left generalized tempered fractional integral is given by
Υ ( x − σ ) − ζ (x− σ )
Z x
ζ
µ1+ TΥ S(x) := e S(σ)dσ, x > µ1 . (13)
µ1 x−σ
Υ ( σ − x ) − ζ ( σ −x)
Z µ2
ζ
µ2− TΥ S(x) := e S(σ)dσ, x < µ2 . (14)
x σ−x
From Definition 4, we can recapture some known integral operators such as Riemann
integrals, Riemann–Liouville fractional integrals, generalized integral operators, and tem-
pered fractional integral operators, and we can present some new integral operators as well.
Some of them are given explicitly in the following Remark 1. Moreover, the importance
of these new integral operators depends on the choices of the parameter ζ, since it takes
values in the domain [0, ∞). That means that interested readers that will do numerical
computations will see clearly the advantage of this new operator compared with other
known operators for suitably choosing this parameter ζ. For ζ = 0, these operators attains
maximum values. For other values of ζ, their values decrease. We believe that these new
operators will open a new door for investigating various variational problems for different
kinds of convexity functions. In our study, we are not focused on numerical analysis.
Remark 1.
(a) Taking ζ = 0 in Definition 4, we have generalized fractional integral operators given from
α
(11) and (12). Choosing Υ(d) = Γd(α) in Definition 4, we get Definition 3.
Fractal Fract. 2023, 7, 579 5 of 19
(b) Choosing Υ(d) = d(µ2 − d)α−1 for α ∈ (0, 1) in Definition 4, we obtain the following
conformable left and right tempered fractional integral operators:
Rx α−1 e−ζ (x−σ) S( σ )dσ,
µ1+ C
α,ζ S(x) :=
µ1 ( µ 2 − x + σ ) x > µ1 ,
µ α−1 e−ζ (σ −x) S( σ )dσ,
α,ζ S(x) : =
R 2
µ− C
2 x ( µ2 + x − σ ) x < µ2 .
Taking Υ(d) = dα exp(−Ad), where A = 1−α α for α ∈ (0, 1) in Definition 4, we have the
following exponential left and right tempered fractional integral operators:
Rx
µ1+ E
α,ζ S(x) := 1
α e−(A+ζ )(x−σ) S(σ)dσ, x > µ1 ,
Rµµ12
µ2− Eα,ζ S(x) := 1
α x e−(A+ζ )(σ−x) S(σ )dσ, x < µ2 .
Furthermore, we define the following expressions that will be used in the sequel.
S(µ1 ) + S(µ2 )
µ1 + µ2 1 h ζ ζ
i
S ≤ µ + T S( µ2 ) + − T S( µ1 ) ≤
Υ µ Υ , (19)
2 2 ΨΥ 1
ζ 2 2
S(u) + S(v)
u+v
S ≤ .
2 2
So, we have
µ1 + µ2 h i
2 ΨΥ
ζ ζ ζ
S ≤ µ1+ TΥ S( µ2 ) + µ2− TΥ S( µ1 ) . (21)
2
This implies that the left-hand side of inequality (19) has been established. In order to
prove the right-hand side of inequality (19), we utilize the fact that S is a convex function
on [µ1 , µ2 ]. This allows us to derive the following inequalities:
and
S((1 − d)µ1 + dµ2 ) ≤ (1 − d)S(µ1 ) + dS(µ2 ). (23)
Adding (22) and (23)), we get
This confirms that the right-hand side of inequality (19) has been established. There-
fore, the proof of Theorem 2 is concluded.
dα
Corollary 1. By taking Υ(d) = in Theorem 2, we obtain the following double inequality for
Γ(α)
convex functions using tempered fractional integral operators:
ζ Θ Υ (S , R)
1 ζ
≤ (27)
4Ψ Υ
∆Υ ∆Υ
ζ ζ
≤ 12 M(µ1 , µ2 ) + 14 N(µ1 , µ2 ) 1 − 2 ,
ΨΥ ΨΥ
ζ ζ
where
M(µ1 , µ2 ) := S(µ1 )R(µ1 ) + S(µ2 )R(µ2 ),
N(µ1 , µ2 ) := S(µ1 )R(µ2 ) + S(µ2 )R(µ1 ),
Fractal Fract. 2023, 7, 579 7 of 19
1 Υ (d)
Z µ2 − µ
ΘΥ (S , R) := e−ζd R(µ2 − d)S(d − µ1 ) + R(d − µ1 )S(µ2 − d) dd,
ζ
0 d
Z µ2 − µ
1 1
∆Υ := Υ(d) e−ζd (µ2 − µ1 − d)dd
ζ
( µ2 − µ1 )2 0
and
µ1 + µ2
2R ≤ R(dµ1 + (1 − d)µ2 ) + R(dµ2 + (1 − d)µ1 ). (29)
2
By multiplying both sides of inequalities (28) and (29), we obtain
µ +µ µ +µ
4S 1 2 2 R 1 2 2
≤ S(dµ1 + (1 − d)µ2 )R(dµ1 + (1 − d)µ2 ) + S((1 − d)µ1 + dµ2 )R(dµ2 + (1 − d)µ1 ) (30)
+S(dµ1 + (1 − d)µ2 )R(dµ2 + (1 − d)µ1 ) + S((1 − d)µ1 + dµ2 )R(dµ1 + (1 − d)µ2 ).
Υ(d(µ2 − µ1 )) −ζ (µ2 −µ1 )d
Multiplying both sides of inequality (30) with e and integrat-
d
ing the resulting inequality with respect to d over [0, 1], we have
+ 0 Υ(d(µd2 −µ1 )) e−ζ (µ2 −µ1 )d S((1 − d)µ1 + dµ2 )R(dµ1 + (1 − d)µ2 )dd.
R1
So, we get
µ +µ µ +µ
4 ΨΥ S 1 2 2 R 1 2 2 ≤ µ+ TΥ S(µ2 )R(µ2 ) + µ− TΥ S(µ1 )R(µ1 )
ζ ζ ζ
R µ2 −µ1 Υ(d) −ζd 1 2
(31)
+ 0 d e S(d − µ1 )R(µ2 − d) + S(µ2 − d)R(d − µ1 ) dd.
This implies that the left-hand side of inequality (27) has been established. In order
to prove the right-hand side of inequality (27), we utilize the fact that S and R are both
convex functions on [µ1 , µ2 ]. This allows us to derive the following inequalities:
So, we get
h i
+T
ζ
S(µ2 )R(µ2 ) +S(µ1 )R(µ1 ) + ΘζΥ (S , R)
ζ
µ2− TΥ
h µ1 Υ i
≤ µ+ TζΥ S(µ2 )R(µ2 ) + µ− TζΥ S(µ1 )R(µ1 ) + 2M(µ1 , µ2 )∆ζΥ (37)
1 2
+N(µ1 , µ2 ) ΨζΥ − 2∆ζΥ .
This implies that the right-hand side of inequality (27) has been established. Therefore,
the proof of Theorem 3 is concluded.
where
1 Υ (d)
Z µ2 − µ
ΘΥ (S) := e−ζd S(d − µ1 )S(µ2 − d)dd
ζ
0 d
and
P( µ1 , µ2 ) : = S 2 ( µ1 ) + S 2 ( µ2 ), Q(µ1 , µ2 ) := 2S(µ1 )S(µ2 ).
dα
Corollary 4. By selecting Υ(d) = in Theorem 3, we derive the following double inequality
Γ(α)
for the product of two convex functions employing tempered fractional integral operators:
Γ(α)
h i
S µ1 +2 µ2 R µ1 +2 µ2 − 4 γ (α,µ −µ ) µ +I
α,ζ S( µ )R( µ ) +
2 2 µ −I
α,ζ S( µ )R( µ )
1 1
ζ 2 1 1 2
≤ 1
4 γζ (α,µ2 −µ1 )
Θα,ζ (S , R) (39)
1 ∆α,ζ 1 ∆α,ζ
≤ 2 M( µ1 , µ2 ) γζ (α,µ2 −µ1 ) + 4 N( µ1 , µ2 ) 1 − 2 γζ (α,µ2 −µ1 ) ,
Fractal Fract. 2023, 7, 579 9 of 19
where
Z µ2 − µ
1
Θα,ζ (S , R) := dα−1 e−ζd S(d − µ1 )R(µ2 − d) + S(µ2 − d)R(d − µ1 ) dd
0
and Z µ2 − µ
1 1
∆ α,ζ
:= dα e−ζd (µ2 − µ1 − d)dd.
( µ2 − µ1 )2 0
( µ2 − µ1 ) R 1
h i (40)
ζ ζ 0
= ζ
2ΨΥ0 ΩΥ (1 − d) − ΩΥ (d) S (dµ1 + (1 − d) µ2 )dd,
ΨΥ
ζ ζ
where Ω Υ (d) and are defined respectively, from (15) and (18). We denote
Z 1h
( µ2 − µ1 ) i
ΩΥ (1 − d) − ΩΥ (d) S 0 (dµ1 + (1 − d)µ2 )dd.
ζ ζ
IS ,Ωζ (µ1 , µ2 ) := ζ
(41)
Υ 2ΨΥ 0
( µ2 − µ1 ) h (1) (2)
i
IS ,Ωζ (µ1 , µ2 ) = I ζ ( µ1 , µ2 ) − I ζ ( µ1 , µ2 ) , (42)
Υ 2ΨΥ
ζ S ,ΩΥ S ,ΩΥ
where Z 1
(1)
ΩΥ (1 − d)S 0 (dµ1 + (1 − d)µ2 )dd
ζ
I ζ ( µ1 , µ2 ) := (43)
S ,ΩΥ 0
and Z 1
(2)
ΩΥ (d)S 0 (dµ1 + (1 − d)µ2 )dd.
ζ
I ζ ( µ1 , µ2 ) := (44)
S ,ΩΥ 0
(1)
I ζ ( µ1 , µ2 )
S ,ΩΥ
1
= ΩζΥ (1 − d) S(dµ1µ+( 1−d) µ2 )
1 − µ2
0
(45)
+ µ1 −1 µ2 0 Υ((1−d1)( µ2 −µ1 )) −ζ (µ2 −µ1 )(1−d)
R1
−d e S(dµ1 + (1 − d)µ2 )dd
= ΨζΥ µS( µ2 )
2 − µ1
− 1
µ2 − µ1 × ζ
µ2− TΥ S( µ1 ).
(2) S(µ1 ) 1
I ζ ( µ1 , µ2 ) = −ΨζΥ + × ζ
µ1+ TΥ S( µ2 ). (46)
S ,ΩΥ µ2 − µ1 µ2 − µ1
Substituting (45) and (46) in (42), we obtain the desired equality (40).
dα
Remark 2. Taking Υ(d) = and ζ = 0 in Lemma 1, we get ([19], Lemma 3.1).
Γ(α)
Fractal Fract. 2023, 7, 579 10 of 19
where p
R1 ζ
ΠΥ (p) := 0 ΩΥ (1 − d) − ΩΥ (d) dd
ζ ζ
R 1h ip R 1h ip (50)
= 02 ΩζΥ (1 − d) − ΩζΥ (d) dd + 1 ΩζΥ (d) − ΩζΥ (1 − d) dd.
2
Proof. Based on the properties of modulus, the convexity of |S 0 |q and applying Hölder’s
inequality, we can deduce from Lemma 1 the following results:
( µ2 − µ1 ) R1
ΩΥ (1 − d) − ΩΥ (d) |S 0 (dµ1 + (1 − d)µ2 )|dd
ζ ζ
IS ,Ωζ (µ1 , µ2 ) ≤ ζ 0
Υ 2ΨΥ
1 1
p p R
≤ (µ2 −ζµ1 ) 0 ΩζΥ (1 − d) − ΩζΥ (d) dd
R1 1 0 (dµ + (1 − d) µ )|q dd q
0 |S 1 2
2ΨΥ
1 1
≤ (µ2 −ζµ1 ) ΠζΥ (p) p 0 d|S 0 (µ1 )|q + (1 − d)|S 0 (µ2 )|q dd
R 1 q
2ΨΥ
" #1
( µ2 − µ1 ) ζ p1 |S 0 (µ2 )|q +|S 0 (µ1 )|q q
= ζ Π Υ (p) 2 .
2ΨΥ
Let us highlight a few specific scenarios that arise as special cases of Theorem 4.
dα
Corollary 5. Taking Υ(d) = and ζ = 0 in Theorem 4, we get ([19], Corollary 3.3).
Γ(α)
( µ2 − µ1 ) p1
Π Υ (p)
ζ
IS ,Ωζ (µ1 , µ2 ) ≤ K ζ
. (51)
Υ 2ΨΥ
Fractal Fract. 2023, 7, 579 11 of 19
(µ −µ )
IS ,Ωζ (µ1 , µ2 ) ≤ 2 ζ 1
2ΨΥ
( Υ
h i1− q1 h i q1
× AζΥ,1 AΥ,2 |S 0 (µ1 )|q + AΥ,1 − AΥ,2 |S 0 (µ2 )|q
ζ ζ ζ
(52)
)
h i1− q1 h i q1
ζ
+ BΥ,1 ζ 0 q ζ ζ 0
BΥ,2 |S (µ1 )| + BΥ,1 − BΥ,2 |S (µ2 )| q ,
where
ζ R 21 h
ζ ζ
i
AΥ,1 := 0ΩΥ (1 − d) − ΩΥ (d) dd,
ζ R1 h ζ ζ
i
AΥ,2 := 02 d ΩΥ (1 − d) − ΩΥ (d) dd,
ζ R 1h ζ ζ
i
BΥ,1 := 1 ΩΥ (d) − ΩΥ (1 − d) dd,
ζ R 21 h ζ ζ
i
BΥ,2 := 1 d ΩΥ (d) − ΩΥ (1 − d) dd
2
ΨΥ
ζ
and is defined from (18).
Proof. From Lemma 1, convexity of |S 0 |q , power mean inequality and properties of the
modulus, we have
IS ,Ωζ (µ1 , µ2 )
Υ
( µ2 − µ1 ) R 1
≤ ζ
ζ ζ
|S 0 (dµ1 + (1 − d)µ2 )|dd
0 Ω Υ (1 − d) − Ω Υ (d)
2ΨΥ
R 1− q1
≤ (µ2 −ζµ1 ) 0 ΩζΥ (1 − d) − ΩζΥ (d) dd
1
2ΨΥ
R q1
1
× 0 ΩζΥ (1 − d) − ΩζΥ (d) |S 0 (dµ1 + (1 − d)µ2 )|q dd
i 1− q1
(
( µ2 − µ1 ) R 12 h ζ ζ
≤ ζ 0 ΩΥ (1 − d) − ΩΥ (d) dd
2ΨΥ
1
R 12 h i q
d|S 0 (µ )|q + (1 − d)|S 0 (µ )|q
ζ ζ
× 0 Ω Υ (1 − d) − Ω Υ (d) 1 2 dd
R h i 1− q1
ζ 1 ζ
+ ΩΥ (d) − ΩΥ (1 − d) dd
1
2 )
R h
1 ζ ζ
i
0 q 0 q
q1
× 1 ΩΥ (d) − ΩΥ (1 − d) d|S (µ1 )| + (1 − d)|S (µ2 )| dd
2
( µ2 − µ1 )
= ζ
2ΨΥ
(
h i1− q1 h i q1
φΥ,2 |S 0 (µ1 )|q + φΥ,1 − φΥ,2 |S 0 (µ2 )|q
ζ ζ ζ ζ
× φΥ,1
)
h i1− q1 h i q1
ζ
+ BΥ,1 ζ 0 q ζ ζ 0
BΥ,2 |S (µ1 )| + BΥ,1 − BΥ,2 |S (µ2 )| q .
Now, let us consider some specific instances that arise as special cases of Theorem 5.
(µ −µ )
IS ,Ωζ (µ1 , µ2 ) ≤ 2 ζ 1
2ΨΥ
( Υ )
h i h i (53)
× AΥ,2 |S 0 (µ1 )| + AΥ,1 − AΥ,2 |S 0 (µ2 )| + BΥ,2 |S 0 (µ1 )| + BΥ,1 − BΥ,2 |S 0 (µ2 )| .
ζ ζ ζ ζ ζ ζ
dα
Corollary 8. Choosing Υ(d) = and ζ = 0 in Corollary 7, we get ([19], Theorem 1.4).
Γ(α)
Corollary 9. By employing the inequality |S 0 | ≤ K in Theorem 5, we can derive the following result,
( µ2 − µ1 ) h ζ ζ
i
IS ,Ωζ (µ1 , µ2 ) ≤ K ζ
AΥ,1 + BΥ,1 . (54)
Υ 2ΨΥ
where Z 1h ip Z 1h ip
ΞΥ,1 (x, p) := ΣΥ,1 (x, d) ΞΥ,2 (x, p) := ΣΥ,2 (x, d)
ζ ζ ζ ζ
dd, dd.
0 0
(x− µ1 ) R1
ΣΥ,1 (x, d)|S 0 (dx + (1 − d)µ1 )|dd
ζ
IS ,Σζ ζ (x; µ1 , µ2 ) ≤ 0
Υ,1 ,ΣΥ,2
ζ
2ΨΥ
( µ2 −x) R 1
ΣΥ,2 (x, d)|S 0 (dx + (1 − d)µ2 )|dd
ζ
+ ζ 0
2Ψ Υ
h ip p1 R 1
(x− µ1 ) R 1 1 0 (dx + (1 − d) µ )|q dd q
Σ
ζ
≤ ζ 0 Υ,1 ( x, d ) dd 0 |S 1
2ΨΥ
h 1 1
i p p R
+ (µ2 −ζx) 0 ΣζΥ,2 (x, d) dd
R1 1 0 (dx + (1 − d) µ )|q dd q
0 |S 2
2ΨΥ
1 R q1
≤ (x−µζ 1 ) ΞζΥ,1 (x, p) p 0 d|S 0 (x)|q + (1 − d)|S 0 (µ1 )|q dd
1
2ΨΥ
1 R 1 q1
+ (µ2 −ζx) ΞζΥ,2 (x, p) p 0 d|S 0 (x)|q + (1 − d)|S 0 (µ2 )|q dd
2ΨΥ
" #1
(x− µ1 ) ζ p1 |S 0 (x)|q +|S 0 (µ1 )|q q
= ζ ΞΥ,1 (x, p) 2
2ΨΥ
" # q1
1 |S 0 (x)|q +|S 0 (µ2 )|q
+ (µ2 −ζx) ΞζΥ,2 (x, p) p
2 .
2ΨΥ
µ1 + µ2
Corollary 10. By substituting x = into Theorem 6, we obtain the following midpoint
2
inequality using generalized tempered fractional integral operators:
µ1 + µ2
( µ2 − µ1 ) ζ p1
IS ,Σζ ζ 2 ; µ1 , µ2 ≤ 1 ζ Ξ Υ (p)
Υ,1 ,ΣΥ,2 2
2+ q
ΨΥ
(56)
q q1 q1 )
(
µ1 + µ2
µ1 + µ2
q
× |S 0 (µ 1 )|q + S0 2 + S0 2 + |S 0 (µ 2 )|q ,
where p
Z d Υ u µ2 − µ1
Z 1
2
Ξ Υ (p) : = e−ζ (µ2 −µ1 )u du dd.
ζ
0 0 u
dα
Corollary 11. Choosing Υ(d) = in Theorem 6, we get
Γ(α)
1
p1
IS ,Σζ ,Σζ (x; µ1 , µ2 ) ≤ 2γζ (α,µ2 −µ1 )
dγ (α, p)
Υ,1 Υ,2
( " # q1 " # q1 ) (57)
|S 0 (x)|q +|S 0 (µ )|q |S 0 (x)|q +|S 0 (µ )|q
× (x − µ 1 ) α +1 2
1
+ ( µ 2 − x ) α +1 2
2
,
where Z 1h ip
dγ (α, p) := γζ (µ2 −µ1 ) (α, d) dd.
0
1 (x− µ )
IS ,Σζ ,Σζ (x; µ1 , µ2 ) ≤ ζ
Υ,1 Υ,2 2ΨΥ
h i1− q1 h i q1
× MζΥ,1 (x) MΥ,3 (x) |S 0 (x)|q + MΥ,1 (x) − MΥ,3 (x) |S 0 (µ1 )|q
ζ ζ ζ
(59)
h i1− q1 h i q1
+ (µ2 −ζx) MζΥ,2 (x) MΥ,4 (x) |S 0 (x)|q + MΥ,2 (x) − MΥ,4 (x) |S 0 (µ2 )|q ,
ζ ζ ζ
2ΨΥ
where R1 ζ
Σ (x, d)dd,
ζ
MΥ,1 (x) :=
R01 ζΥ,1
MΥ,2 (x) := 0 ΣΥ,2 (x, d)dd,
ζ
ζ R1 ζ
MΥ,3 (x) := 0 dΣΥ,1 (x, d)dd,
ζ R1 ζ
MΥ,4 (x) := 0 dΣΥ,2 (x, d)dd
Proof. From Lemma 2, convexity of |S 0 |q , power mean inequality and properties of the
modulus, we have
(x− µ1 ) R1
ΣΥ,1 (x, d)|S 0 (dx + (1 − d)µ1 )|dd
ζ
IS ,Σζ ζ (x; µ1 , µ2 ) ≤ 0
Υ,1 ,ΣΥ,2
ζ
2ΨΥ
( µ2 −x) R 1
ΣΥ,2 (x, d)|S 0 (dx + (1 − d)µ2 )|dd
ζ
+ ζ 0
2Ψ Υ
R 1− 1 R 1
≤ (x−µζ 1 ) 0 ΣζΥ,1 (x, d)dd
1 q 1 ζ 0 (dx + (1 − d) µ )|q dd q
Σ
0 Υ,1 ( x, d )|S 1
2ΨΥ
R 1− 1 R 1
+ (µ2 −ζx) 0 ΣζΥ,2 (x, d)dd
1 q 1 ζ q
0 ΣΥ,2 (x, d)|S 0 (dx + (1 − d)µ2 )|q dd
2ΨΥ
h i1− q1 R 1
≤ (x−µζ 1 ) MζΥ,1 (x)
1 ζ 0 (x)|q + (1 − d)|S 0 ( µ )|q dd q
Σ
0 Υ,1 ( x, d ) d |S 1
2ΨΥ
h i1− q1 R 1
+ (µ2 −ζx) MζΥ,2 (x)
1 ζ 0 (x)|q + (1 − d)|S 0 ( µ )|q dd q
Σ
0 Υ,2 ( x, d ) d |S 2
2ΨΥ
h i1− q1 h i q1
= (x−µζ 1 ) MζΥ,1 (x) MΥ,3 (x) |S 0 (x)|q + MΥ,1 (x) − MΥ,3 (x) |S 0 (µ1 )|q
ζ ζ ζ
2ΨΥ
h i1− q1 h i q1
+ (µ2 −ζx) MζΥ,2 (x) MΥ,4 (x) |S 0 (x)|q + MΥ,2 (x) − MΥ,4 (x) |S 0 (µ2 )|q .
ζ ζ ζ
2ΨΥ
1 (x− µ )
IS ,Σζ ,Σζ (x; µ1 , µ2 ) ≤ ζ
h Υ,1 Υ,2 2Ψ Υ i
× MζΥ,3 (x) |S 0 (x)| + MζΥ,1 (x) − MζΥ,3 (x) |S 0 (µ1 )| (60)
h i
+ (µ2 −ζx) MζΥ,4 (x) |S 0 (x)| + MζΥ,2 (x) − MζΥ,4 (x) |S 0 (µ2 )| .
2ΨΥ
µ1 + µ2
Corollary 14. Choosing x = in Theorem 7 and utilizing the generalized tempered frac-
2
tional integral operators, we obtain the following midpoint inequality:
i 1
ζ 1− q
h
µ1 + µ2 ( µ2 − µ1 )
IS ,Σζ ζ 2 ; µ1 , µ2 ≤ MΥ
Υ,1 ,ΣΥ,2
ζ
4ΨΥ
( q1
µ1 + µ2
q
× ζ
NΥ S0 2 + ζ
MΥ − NζΥ |S 0 (µ 1 )|q (61)
1 )
µ1 + µ2
q q
+ NζΥ S 0 2 + MζΥ − NζΥ |S 0 (µ2 )|q ,
where !
µ −µ
R1 Rd Υ u 22 1
e−ζ (µ2 −µ1 )u du
ζ
MΥ := 0 0 u dd,
!
R d Υ u µ2 −2 µ1
R1
e−ζ (µ2 −µ1 )u du dd.
ζ
NΥ : = 0 d 0 u
Fractal Fract. 2023, 7, 579 15 of 19
dα
Corollary 15. Taking Υ(d) = in Theorem 7, we obtain
Γ(α)
h i1− q1
IS ,Σζ ,Σζ (x; µ1 , µ2 ) ≤ 2γ (α,µ1 −µ ) dγ (α)
2 1
Υ,1 Υ,2 h
ζ
i q1
× (x − µ1 )α+1 dγ,1 (α) |S 0 (x)|q + dγ (α) − dγ,1 (α) |S 0 (µ1 )|q
(62)
h i q1
0
+ 1 q
0 q
+(µ2 − x) α dγ,1 (α) |S (x)| + dγ (α) − dγ,1 (α) |S (µ2 )| ,
where R1
dγ ( α ) : =0R γζ (µ2 −µ1 ) ( α, d)dd,
1
dγ,1 (α) := 0 d γζ (µ2 −µ1 ) (α, d)dd.
K n o
IS ,Σζ ζ (x; µ1 , µ2 ) ≤ (x − µ1 )MζΥ,1 (x) + (µ2 − x)MζΥ,2 (x) . (63)
Υ,1 ,ΣΥ,2
ζ
2ΨΥ
Remark 3. Applying our results for suitable choices of function Υ(d) = d(µ2 − d)α−1 and
Υ(d) = dα exp(−Ad), where A := 1−α α for α ∈ (0, 1) such that |S 0 |q to be convex function, we
can construct some new tempered fractional integral type inequalities. We omit their proofs and the
details are left to the interested readers.
4. Applications
In the final section, we provide several examples that illustrate the results we have
established in relation to matrices, modified Bessel functions and q-digamma function.
4.1. Matrices
Example 1. In this context, the set of n × n complex matrices is denoted by Cn . Similarly, Mn represents
the algebra of n × n complex matrices and Mn+ refers to the subset of strictly positive matrices within
Mn . Consequently, for a matrix A ∈ Mn+ , it holds that hAu, ui > 0 for all nonzero u ∈ Cn .
Incorporating the concepts of matrices and convexity, Sababheh [27] introduced the function
S(d) := kAd XB1−d + A1−d XBd k, where A, B ∈ Mn+ , and X ∈ Mn . It was shown that this
function is convex for all d ∈ [0, 1]. Consequently, by utilizing Theorem 2 for any 0 ≤ µ1 < µ2 ≤ 1,
we obtain
µ1 + µ2 µ1 + µ2 µ1 + µ2 µ1 + µ2
kA 1− + A1− k
h XB XB
2 2 2 2
i
≤ 1 ζ 1−µ2 + A1−µ2 XBµ2 k + ζ 1− µ1 + A1−µ1 XBµ1 k
+ T kA 2 XB
µ2− TΥ kA XB
µ µ1
(64)
2 Ψ Υ µ1 Υ
ζ
a ρ+2n
dρ (a ) = ∑ 2
n!Γ(n + ρ + 1)
,
n≥0
π d− ρ (a ) − dρ (a )
Bρ (a ) : = .
2 sin ρπ
Fractal Fract. 2023, 7, 579 16 of 19
Jρ (a ) : = 2ρ Γ ( ρ + 1 )a− ρ Bρ (a ),
where Γ(·) represents the gamma function.
In the cited reference [28], two derivative formulas for Jρ (a) are given. The first
derivative is expressed as:
a
J0ρ (a) = J (a). (65)
2 ( ρ + 1 ) ρ +1
The second derivative is given by:
a2 Jρ+2 (a) Jρ +1 (a )
J00ρ (a) = + . (66)
4(ρ + 1)(ρ + 2) 2(ρ + 1)
Example 2. Applying Corollary 7 and using expressions (65) and (66), we have
h i
1 ζ ζ ( µ2 − µ1 )
µ 1 Jρ +1 ( µ 1 ) + µ 2 Jρ +1 ( µ 2 ) − µ2 · µ1+ TΥ Jρ+1 ( µ2 ) + µ1 · µ2− TΥ Jρ+1 ( µ1 ) ≤
ΨΥ
ζ ζ
2ΨΥ
(
µ21 µ22
ζ
× AΥ,2 ρ+2 Jρ+2 ( µ1 ) + 2Jρ+1 ( µ1 ) + AζΥ,1 − AζΥ,2 ρ+2 Jρ+2 ( µ2 ) + 2Jρ+1 ( µ2 )
)
µ2 µ22
ζ ζ
+ BΥ,2 ρ+12 Jρ+2 (µ1 ) + 2Jρ+1 (µ1 ) + BΥ,1 − BζΥ,2 ρ+2 Jρ+2 ( µ2 ) + 2Jρ+1 ( µ2 ) .
Proof. By utilizing Corollary 7 and incorporating Equations (65) and (66) where the func-
tion S(a) := J0ρ (a) with a > 0, we can derive the final result, thus completing the proof.
Example 3. Using Corollary 13 and applying expressions (65) and (66), we get
h i
1 ζ ζ
2xJρ+1 (x) − µ2 · x+ TΥ Jρ +1 ( µ 2 ) + µ1 · x− TΥ Jρ +1 ( µ 1 )
ΨΥ
ζ
"
(x− µ1 ) ζ x2
≤ ζ MΥ,3 (x) ρ+2 Jρ+2 (x) + 2Jρ+1 (x)
2ΨΥ
#
µ21
ζ ζ
+ MΥ,1 (x) − MΥ,3 (x) ρ+2 Jρ+2 ( µ1 ) + 2Jρ+1 ( µ1 )
"
+ (µ2 −ζx) ζ
MΥ,4 (x) x2
ρ+2 Jρ+2 (x) + 2Jρ+1 (x)
2ΨΥ
#
µ22
ζ ζ
+ MΥ,2 (x) − MΥ,4 (x) ρ+2 Jρ+2 ( µ2 ) + 2Jρ+1 ( µ2 ) .
Proof. By employing Corollary 13 and incorporating Equations (65) and (66), where the function
S(a) := J0ρ (a) with a > 0, we can deduce the outcome, thereby concluding the proof.
or alternatively,
Fractal Fract. 2023, 7, 579 17 of 19
∞
qkγ
$q (γ) = − ln(1 − q) + ln q ∑ kγ
.
k=0 1 − q
This expression holds true when q > 1 and γ > 0. The q-digamma function $q can
also be expressed in the following form:
∞
" #
1 q−(k+γ)
$q (γ) = − ln(q − 1) + ln q γ − − ∑
2 k=0 1 − q−(k+γ)
or alternatively,
∞
# "
1 q−kγ
$q (γ) = − ln(q − 1) + ln q γ − − ∑ .
2 k=0 1 − q−kγ
Proof. The statement can be readily derived by utilizing Corollary 7, where S(a) → $q (a)
represents a completely monotone function on the interval (0, ∞) for all a > 0. As a result,
S(a) := $q0 (a) is a convex function.
2ΨΥ
(x− µ1 ) ζ 00 (x) + Mζ (x) − Mζ (x) $00 (µ )
≤ ζ M Υ,3 ( x ) $ q Υ,1 Υ,3 q 1
2ΨΥ
+ (µ2 −ζx) MζΥ,4 (x) $q00 (x) + MζΥ,2 (x) − MζΥ,4 (x) $q00 (µ2 ) .
2ΨΥ
Proof. The assertion can be immediately derived by utilizing Corollary 13, where S(a) →
$q (a) represents a completely monotone function on the interval (0, ∞) for all a > 0.
Consequently, S(a) := $q0 (a) is a convex function.
5. Conclusions
Overall, this paper aimed to contribute to the expanding field of fractional H–H
inequalities by presenting new results, exploring their properties, investigating their con-
nections with fractional calculus, and demonstrating their applications. Through our work,
we sought to advance the theoretical foundations of fractional calculus, enhance our un-
derstanding of convex functions in the fractional calculus framework, and inspire further
research in this exciting and evolving area of mathematics.
This research paper introduces novel concepts of left and right generalized tempered
fractional integral operators and establishes fresh H–H inequalities for convex functions
and their products. Additionally, it derives two useful identities for differentiable functions
that involve the generalized tempered fractional integral operator. These identities are
then utilized to establish H–H and midpoint-type integral inequalities for convex functions.
The paper also explores various special cases and demonstrates how the general findings
recover known results. Furthermore, the paper presents compelling applications related to
matrices, modified Bessel functions and q-digamma functions. By employing the newly
introduced generalized tempered fractional integral operators, along with well-known
Fractal Fract. 2023, 7, 579 18 of 19
Author Contributions: Conceptualization, A.K., S.K.S., A.M.M. and Y.A.; methodology, A.K., S.K.S.
and A.M.M.; software, S.K.S. and Y.A.; validation, A.K., A.M.M. and S.K.S.; project administration,
S.K.S.; writing—original draft preparation, A.K. and S.K.S.; writing—review and editing, A.K., S.K.S.,
A.M.M. and Y.A. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Not applicable.
Acknowledgments: The authors extend their appreciation to the Deanship of Scientific Research at
King Khalid University for funding this work through the large group research project under grant
number RGP2/366/44.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Hadamard, J. Étude sur les propriétés des fonctions entières en particulier d’une fonction considérée par Riemann. J. Math. Pures
Appl. 1893, 58, 171–215.
2. Bayraktar, B.; Kórus, P.; Valdés, J.E.N. Some new Jensen-Mercer type integral inequalities via fractional operators. Axioms 2023,
12, 517.
3. Sahoo, S.K.; Jarad, F.; Kodamasingh, B.; Kashuri, A. Hermite-Hadamard type inclusions via generalized Atangana-Baleanu
fractional operator with application. AIMS Math. 2022, 7, 12303–12321.
4. Tariq, M.; Ahmad, H.; Sahoo, S.K.; Kashuri, A.; Nofal, T.A.; Hsu, C.H. Inequalities of Simpson-Mercer-type including Atangana-
Baleanu fractional operators and their applications. AIMS Math. 2022, 7, 15159–15181.
5. Kashuri, A.; Sahoo, S.K.; Mohammed, P.O.; Sarairah, E.A.; Hamed, Y.S. Some new Hermite-Hadamard type inequalities pertaining
to fractional integrals with an exponential kernel for subadditive functions. Symmetry 2023, 15, 748.
6. Bayraktar, B.; Nápoles, J.E.; Rabossi, F. Some refinements of the Hermite–Hadamard inequality with the help of weighted integrals.
Ukr. Mat. Zhurnal 2023, 75, 736–752.
7. Aslani, S.M.; Delavar, M.R; Vaezpour, S.M. Inequalities of Fejér type related to generalized convex functions with applications.
Int. J. Anal. Appl. 2018, 16, 38–49.
8. Delavar, M.R.; Dragomir, S.S. On η-convexity. Math. Inequal. Appl. 2017, 20, 203–216.
9. Delavar, M.R.; De La Sen, M. Some generalizations of Hermite-Hadamard type inequalities. SpringerPlus 2016, 5, 1661.
10. Dragomir, S.S.; Agarwal, R.P. Two inequalities for differentiable mappings and applications to special means of real numbers and
trapezoidal formula. Appl. Math. Lett. 1998, 11, 91–95.
11. Jleli, M.; Samet, B. On Hermite-Hadamard type inequalities via fractional integral of a function with respect to another function.
J. Nonlinear Sci. Appl. 2016, 9, 1252–1260.
12. Kashuri, A.; Liko, R. Some new Hermite-Hadamard type inequalities and their applications. Stud. Sci. Math. Hung. 2019,
56, 103–142.
13. Liu, W.; Wen, W.; Park, J. Hermite-Hadamard type inequalities for MT-convex functions via classical integrals and fractional
integrals. J. Nonlinear Sci. Appl. 2016, 9, 766–777.
14. Mihai, M.V. Some Hermite-Hadamard type inequalities via Riemann-Liouville fractional calculus. Tamkang J. Math. 2013,
44, 411–416.
15. Omotoyinbo, O.; Mogbodemu, A. Some new Hermite-Hadamard integral inequalities for convex functions. Int. J. Sci. Innov.
Technol. 2014, 1, 1–12.
16. Özdemir, M.E.; Dragomir, S.S.; Yildiz, C. The Hadamard’s inequality for convex function via fractional integrals. Acta Math. Sci.
2013, 33, 153–164.
17. Qi, F.; Mohammed, P.O.; Yao, J.C.; Yao, Y.H. Generalized fractional integral inequalities of Hermite-Hadamard type for (α, m)-
convex functions. J. Inequal. Appl. 2019, 135, 17.
18. Sarikaya, M.Z.; Ertuğral, F. On the Generalized Hermite-Hadamard Inequalities. 2017, p. 20. Available online: https://www.
researchgate.net/publication/321760443 (accessed on 21 June 2023).
Fractal Fract. 2023, 7, 579 19 of 19
19. Set, E.; Gözpinar, A. A study On Hermite-Hadamard type inequalities for s-convex functions via conformable fractional integrals.
Stud. Univ. Babeş-Bolyai Math. 2017, 62, 309–323.
20. Set, E.; Noor, M.A.; Awan, M.U.; Gözpinar, A. Generalized Hermite-Hadamard type inequalities involving fractional integral
operators. J. Inequal. Appl. 2017, 169, 10.
21. Xi, B.Y.; Qi, F. Some integral inequalities of Hermite-Hadamard type for convex functions with applications to means. J. Funct.
Spaces Appl. 2012, 2012, 980438.
22. Mohammed, P.O.; Sarikaya, M.Z.; Baleanu, D. On the generalized Hermite-Hadamard inequalities via the tempered fractional
integrals. Symmetry 2020, 12, 595.
23. Sarikaya, M.Z.; Yildirim, H. On generalization of the Riesz potential. Indian J. Math. Math. Sci. 2007, 3, 231–235.
24. Katugampola, U.N. New approach to a generalized fractional integral. Appl. Math. Comput. 2011, 218, 860–865.
25. Fahad, A.; Butt, S.I.; Bayraktar, B.; Anwar, M.; Wang, Y. Some new Bullen-type inequalities obtained via fractional integral
operators. Axioms 2023, 12, 691.
26. Latif, M.A.; Kalsoom, H.; Khan, Z.A.; Al-moneef, A.A. Refinement mappings related to Hermite-Hadamard type inequalities for
GA-convex function. Mathematics 2022, 10, 1398.
27. Sababheh, M. Convex functions and means of matrices. Math. Inequal. Appl. 2017, 20, 29–47.
28. Watson, G.N. A Treatise on the Theory of Bessel Functions; Cambridge University Press: Cambridge, UK, 1995.
29. Jain, S.; Mehrez, K.; Baleanu, D.; Agarwal, P. Certain Hermite-Hadamard inequalities for logarithmically convex functions with
applications. Mathematics 2019, 7, 163.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.