1. Introduction
Convex analysis based on the convex set and function whose epigraph is convex. It is an extensively investigated aspect due to some distinctive and unique properties regarding minima and maxima. As these concepts are purely geometric properties of sets and functions, they provide more information to enquire about numerous problems in linear programming and optimization. Using the idea of weighted means, different generalizations, refinements, and extensions have been created for different problems. These include weighted arithmetic means, harmonic means, geometric means, and other means. These have led to the classical convex set, the harmonic set, and the geometric set. Also, it is worth mentioning that non-convex sets and convex functions are effective tools to tackle non-convex sets, and functions introduced on these kinds of sets enjoy similar properties as convex functions.
The impact of convexity is unforgettable in various domains of mathematics, such as advanced analysis, topological spaces, particularly separation axioms, fixed point theory, and optimization. The convexity of the functions can be investigated through derivatives of the functions; a positive second-order derivative determines the concavity of functions.
The theory of inequalities covers many disciplines of mathematical analysis and other subjects due to its applicable characteristics. Among them, one of the applications of integral inequalities is to conclude some more refined upper bounds of error estimations of numerical quadrature rules like the trapezoidal rule, midpoint rule, Ostrowski’s rule, Simpson’s rules, etc. involving convex functions and their generalizations. Also, these bounds provide some relations between special means, special functions, probability theory, information theory, etc. By applying the notion of convexity, various fundamental inequalities can be extracted, like Jensen inequality,
-
inequality, etc. From the following perspective, we revisit the familiar result due to Hermite and Hadamard separately: Let
be a convex function, then:
This inequality can be viewed as another criterion of convex functions. For further detail, consult [
1,
2]. In 2019, Wu et al. [
3] purported the unified form of convexity, which is described as follows.
Definition 1 ([
3]).
Let be continuous monotonic function. Then, is considered to be ⋏-convex set based on ⋏ if: Now, we revisit the class ⋏-convex function, which is defined in [
3].
Definition 2 ([
3]).
A function is said to be ⋏-convex function with respect to strictly monotonic continuous function ⋏ if: Another straightforward question in the realm of research is to formulate the set-valued counterparts of the known and new problems. An interval-valued function is a special type of set-valued function with co-domain as the collection of all real positive intervals. Moore was the first person to study the forgotten subject of interval analysis in a very systematic approach to derive the error estimate of finite machines. For more detail, see the monographs published by Moore [
4]. The remarkable work of Moore is regarded as a re-emergence of this subject. After this, many authors have implemented
concepts together with fractional calculus to study the various applicable fields of sciences. Working in the following directions, Breckner [
5] delivered the conception of set-valued convex functions, which is demonstrated below.
Definition 3. Let is said to be convex function, if: specifies the collection of all real intervals. For the first time, Sadowska [
6] initiated the idea to extend the inequalities by set-valued functions and he investigated the
-
inequality considering the set-valued convex function. Which is described below.
Let
be an
convex function, then:
If
is a collection of all divisions of
and
be the collection of all divisions
P such that mesh
, then
is referred to as interval-valued Riemann integrable on
, if ∃
and for each
there exit
such that:
where
specifies the Riemann sum of
for any
. The expression (
1) represents that
is the
-integral of
such that:
For the sake of brevity, we specify the space of Riemann integrable functions and Riemann integration on and by and , respectively.
Theorem 1 ([
4]).
Let be an function such that:and The space of Lebesgue integrable functions is defined by
. Now, we recover the R-L integral operators, which are provided in [
7].
Definition 4. Let , then: Analogously, the right side of the R-L operator is given as:where is the gamma function. Now, we reproduce the R-L operators.
Definition 5 ([
8]).
Let be an function such that , then:andwith . Obviously, we observe that:and Nowadays, several integral inequalities are generalized and rectified by implementing
functions based on different partial and total ordered relations. Chalco-Cano et al. [
9,
10] analyzed the error inequality of the rectangular quadrature rule by making use of the generalized Hukuhara difference and interval-valued functions and also provided some useful applications, respectively.
Furthermore, Costa et al. [
11] developed some interesting inequalities by considering the fuzzy real-valued functions. These papers were the initial attempts to make the inequalities more charming from the implementation aspect. Still, the above-mentioned studies are the gateway for further progress in the field of mathematical inequalities, especially those associated with set-valued functions. In the following scenario, Flores et al. [
12] proposed the novel counterparts of renowned inequalities involving generalized functions. Sharma et al. [
13] established the integral inequalities of trapezium like through non-convex
functions defined over invex sets. In [
14,
15], Zhao and their coauthors investigated and explored the renowned Jensen’s and
-
-like containments by utilizing the
unified convexity, which yields several notions of convexity through appropriate substitutions and Chebyshev-type inclusions, respectively. In [
16], Abdeljawad et al. have derived some more general fractional containments of Hermite-Hadamar’s type inequalities involving
p-
convexity. Khan et al. [
17] gave the conception of fuzzy convex functions for a coordinate system and applied it to construct fresh coordinated trapezium-type inclusions. Fractional calculus has valuable impacts on the rapid progression of several disciplines, including approximation theory and inequalities. For recent developments in fractional calculus and approximation theory, see [
18,
19,
20,
21,
22,
23].
Sarikaya et al. [
24] started the use of non-integer order calculus to find fractional analogues of
-
inequality considering R-L fractional operators. In [
25], the authors explored the interval-valued unified fractional operator and concluded some fractional parametric integral inclusions.
Mohsin and their coauthors [
26] explored two-dimensional
harmonic convex functions and general family operators based on Raina’s function to develop new
-
-like inequalities. In [
27], the authors formulated the fractional tempered trapzoidal-like inequalities and discussed the implications of primary findings. In [
28], Akdemir and their coauthors calculated the Chebyshev-type formulas via a general family of fractional operators. Set et al. [
29] explored the fractional versions of inequalities through Atangana-Baleanu fractional operators and the convexity property of the functions. In [
8], Budak and their colleagues examined the fractional trapezium type inequalities through
convex functions. In 2022, Du and Zhou [
30] examined the coordinated set-valued variants of Hermite–Hadamard-like inequalities associated with fractional integral operators having non-singular kernels. Kara et al. [
31] explored the two-dimensional interval-valued integral inequalities by means of unified fractional operators.
In [
32], the authors implemented the technique of
convex functions and post-quantum calculus to examine some fresh representations of already established findings. In [
33], the authors investigated the non-convex functions based on
connected sets in fuzzy domains and delivered some variants of trapezoid-type inequalities. Cortez et al. [
34] gave the idea of a generic class of convexity and computed new counterparts of renowned Jensen’s and
-
type inequalities along with applications.
Motivated by the above-mentioned studies, we aim to unify the existing notion of convexity in interval analysis. In the subsequent viewpoints, we introduced the idea of - convex functions through the utilization of weighted arithmetic means involving a strictly monotone function ⋏ and non-negative function ℏ. In the setting of some suitable substitutions ⋏ and ℏ, it yields already-known notions and several new classes of convexity, which are discussed in the main section. Moreover, we provide the characterization of this class through inequalities. As applications of this class, we will derive some generic Jensen’s inequality, fractional variants of reverse Minkowski inequality, Hölders inequality, - inequality, its weighted form known as Fejer-- inequality, and some inequalities for the product of functions that are recognized as Pachpatee’s type inclusions. The novelty of the current proceeding is the generic class of convexity and its consequences in inequalities, because it is a larger space of functions containing convex and non-convex function classes. From our constructed results, one can characterize extensive function classes. Moreover, our results will be powerful tools for the computation of various bounds for R-L fractional operators. To check the correctness of the proposed results, some simulations for numerical examples are listed. We hope this study helps curious readers to prove some other sort of inequalities and related problems in optimization.
2. Main Results
In this section, we present our main results.
2.1. Fractional Reverse Minkowski and Hölder’s Inequality
Now, we present the reverse fractional Minkowski’s integral inequality.
Theorem 2. Let be interval-valued functions such that , , and , then:where , , and with . Proof. Since
, then:
Multiplying
on both sides of (
2) and then taking the integration with respect to
over
, we have:
Now, by comparing the above expression with the R-L fractional operator, we obtain:
Also,
, then one can write:
Multiplying
on both sides of (
4) and then taking the integration with respect to
over
, then:
Furthermore, we note that:
From
, we achieve:
Multiplying
on both sides of (
7) and then taking the integration with respect to
over
, then:
Now, from
, we obtain:
Multiplying
on both sides of (
9) and taking the integration with respect to
over
, then:
From (
8) and (
10), we have:
Taking the sum of (
6) and (
11), we acquired our desired result. □
Now, we present the reverse fractional Hölder’s integral inequality.
Theorem 3. Let be interval-valued functions such that , , and , then:where , , and , with and . Proof. Since
, then:
Multiplying
on both sides of (
12) and then taking the integration with respect to
over
, then:
Analogously, by the virtue of
, we have:
Multiplying
on both sides of (
14) and then taking the integration with respect to
over
, then:
Also, by combining (
13) and (
15), we have:
By similar proceedings, we acquire the following containment for
, as:
This completes the proof. □
2.2. On Interval-Valued Convex Functions and Related Results
Now, we initiate the concept of - convex functions, which shows that the several existing classes of convexity and various new generalizations of convexity can be obtained as special cases.
Definition 6. Let be an - function satisfying the condition and be a non-negative function. Then: and .
Now, we report interesting outcomes concerning Definition 6.
Choosing
, we obtain the ⋏-
convex function:
Choosing
and
. Then, we attain the
harmonic convex function defined in [
35]:
Choosing
and
. Then, we attain the
-
p convex functions, which are defined in [
16]:
Remark 1. By fixing , we recover the definition of convex functions.
Definition 7. By fixing in Definition 6, we acquire the convex function: Now, we report the primary deductions of Definition 7.
By specifying , we obtain harmonically s convex functions, which are as follows: Setting , , then we obtain - convex functions, which are as follows:
Definition 8. Choosing in Definition 6, we attain the Godunova–Levin type convex function: For the different choices of , we attain some interesting classes of convexity.
Choosing
, we achieve the
Godunove–Levin
s convex function:
Choosing
. Then, we attain the
Godunova–Levin harmonic
s convex functions:
Choosing
. Then, we attain the
Godunova–Levin
convex functions:
Definition 9. Choosing in Definition 6, we attain - convex function: For the different choices of , we attain some interesting classes of convexity.
For our own ease, the collection of - convex functions, - concave functions, convex functions, and concave functions are represented by , , , and , respectively.
Theorem 4. Let be an function such that with . Then, .
Proof. Assume that
and
and
, then:
Equations (
20) and (
21) indicate that
and
. Conversely, suppose that
and
, then:
and
Hence, the result is achieved. □
Now, we establish Jensen’s inequality in a general sense.
Theorem 5. Let , then:for and . Proof. To acquire our result, we utilize the technique of mathematical induction. Suppose
in (
24), then:
Presume that the inclusion (
24) is satisfied for
such that:
Now, we show that it is true for
.
Hence, the result is achieved. □
Now, we report some consequences of Theorem 5.
Substituting
. Then, we obtain Jensen’s inequality for
h-convex functions, which is derived in [
14].
Substituting
. Then, we obtain Jensen’s inequality for harmonically
h-convex functions:
Substituting
,
. Then, we obtain Jensen’s inequality for
p-convex functions:
Substituting
. Then, we obtain Jensen’s inequality for
-
h convex functions:
Now, we establish the unified - inequality.
Theorem 6. Let , then: Proof. Since
is
-
convex function, then:
Substituting
and
in the above inclusion and then taking the product on both sides by
and then applying the integration with respect to
on
:
Combining (
26) and (
27), we obtain:
Again, employing
-
convexity of
, we have
and
Adding (
28) and (
29), taking the product on both sides by
, and applying the integration with respect to
on
, we acquired our required relation. □
Now, we provide novel results concerning with Theorem 6.
Choosing
in Theorem 6, we obtain:
Choosing
in Theorem 6, we attain fractional
-
inequality for
convex functions, which is proved in [
8]:
Choosing
,
in Theorem 6, then:
Choosing
,
and
in Theorem 6, we obtain:
where
.
Example 1. Presume each of the premises of Theorem 6 are met, we consider with and within , then: This example confirms the accuracy of the --Fejer inequality for the product of - convex functions and symmetric functions. By considering these aforementioned computations, one can easily analyze bounds for R-L operator with different symmetric functions as a kernel.
For the pictorial demonstration, we vary .
Figure 1 reflects the comparison between the left, middle and right sides of Theorem 1. Furthermore red, blue and green color shows lower and upper functions of left, middle and right sides respectively.
Now, we present a unified fractional version of Fejer-Type -’s inequality.
Theorem 7. Let , and be a symmetric function with respect , then: and . Proof. Since
is
-
convex function, for
, then we have:
Substituting
and
in the above relation, then multiplying both sides by
:
By taking the benefit of the fact
:
In this way, we attain our desired inclusion.
To achieve our second inclusion, we take the sum of (
28) and (
29), then taking the product on both sides by
and integration with respect to
on
, then:
Some simple calculations result in the desired inclusion. □
We extract important results concerning Theorem 7.
Choosing
in Theorem 7, we have:
where
and
.
Choosing
and
in Theorem 7, then:
where
and
.
Choosing
in Theorem 7, then:
where
and
.
Choosing
and
in Theorem 7, then:
where
and
.
Example 2. Presume each of the premises of Theorem 7 are met, we consider , where with and symmetric function with respect to such that and is defined as: This example confirm the accuracy of - inequality for the product of - convex functions. By considering these aforementioned computations, one can easily analyze bounds for R-L operator.
Figure 2 reflects the comparison between the left, middle and right sides of Theorem 7. Furthermore red, blue and green color shows lower and upper functions of left, middle and right sides respectively.
Theorem 8. Let :where and . Proof. Since
, then:
Multiplying (
35) and (36), we have:
By adding (
37) and (
38), taking the product of obtained result with
and then applying the integration with respect to
on
:
Now, we utilize the R-L fractional operators, then:
This completes the proof. □
Now, we exhibit interesting results concerning Theorem 8.
Choosing
in Theorem 8, then:
Choosing
and
in Theorem 8, then:
Choosing
in Theorem 8, then:
Choosing
and
in Theorem 8, then:
Example 3. Presume each of the premises of Theorem 8 are met, we consider , where and with , then: The aforementioned computations justify the accuracy of right Pachpatte’s type inequality for the product two - convex functions. By considering these calculations, one can easily evaluate the lower bounds for the R-L operator.
For pictorial depiction, we vary
.
Figure 3 reflects the comparison between the left and right sides of Theorem 8. Furthermore purple and green color shows lower and upper functions of left and right sides respectively.
Theorem 9. Let , then:where and are given by (33) and (34), respectively. Proof. Let
, then:
Multiplying preceding inclusion by
and applying integration with respect to
on
, then:
Employing the definition of
convexity, we acquire:
From (
40)–(
42), we secured our result. □
Now, we explore some crucial results concerning with Theorem 9.
Choosing
in Theorem 9, then:
where
and
are given by (
33) and (34), respectively.
Choosing
and
in Theorem 9, then:
where
and
are given by (
33) and (34), respectively.
Choosing
and
in Theorem 9, then:
where
and
are given by (
33) and (34), respectively.
Example 4. Presume each of the premises of Theorem 9 are met, we consider where and with , then: The aforementioned computations justify the accuracy of Theorem 9 for the product two - convex functions. Moreover, from these computations, one can easily calculate the upper bounds of the R-L operator.
For a pictorial depiction, we vary
.
Figure 4 reflects the comparison between both sides of Theorem 9. Furthermore purple and green color show lower and upper functions of left and right sides respectively.