Strongly Reciprocally (p, h)-Convex Functions

And Some Inequalities

By
Farah Jamil
(Registration No: 00000402935)

Supervisor: Dr. Matloob Anwar

A thesis submitted in partial fulfillment of the requirements

for the degree of Master of Science in Mathematics

School of Natural Sciences

National University of Sciences and Technology (NUST)

H-12, Islamabad, Pakistan

(2024)
Strongly Reciprocally (p, h)-Convex Functions
And Some Inequalities

By
Farah Jamil
(Registration No: 00000402935)

Supervisor: Dr. Matloob Anwar

A thesis submitted in partial fulfillment of the requirements

for the degree of Master of Science in Mathematics

School of Natural Sciences

National University of Sciences and Technology (NUST)

H-12, Islamabad, Pakistan

(2024)
 ACKNOWLEDGEMENTS

I am grateful to Allah Almighty, Who is the Most Merciful and Beneficent, Creator of
the entire universe. I am thankful for His countless blessings, which bestowed upon me
the ability and strength to complete my thesis successfully.
My sincere gratitude goes out to my supervisor, Dr. Matloob Anwar, your exceptional
mentorship and encouragement have made a significant impact on my academic journey.
Without your guidance, I couldn’t achieve my research objective. I am also grateful
to my G.E.C. members, Dr. Muhammad Ishaq and Hafiz Muhammad Fahad, for their
contribution in reviewing my research work.
I extend my gratitude to everyone, especially my parents, brothers, sisters, daughter,
friends, and teachers who helped me, appreciate my efforts, and fueled my motivation
in the whole journey, to shape my research.

I
Contents

LIST OF FIGURES IV

LIST OF SYMBOLS, ABBREVIATIONS AND ACRONYMS V

ABSTRACT 1

1 Introductioni 3

1.1 Convex Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2 Convex Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.2.1 Properties of Convex Function . . . . . . . . . . . . . . . . . . . 7

1.3 Continuity and Differentiability of Convex functions . . . . . . . . . . . 8

1.3.1 Boundedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.3.2 Continuity andidifferentiability . . . . . . . . . . . . . . . . . . 9

1.3.3 Convex functions: Applications, Extensions, and Generalizations 12

1.4 Some remarkable Inequalities . . . . . . . . . . . . . . . . . . . . . . . 13

1.4.1 Jensen’siinequality . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.4.2 Hermite-Hadamardiinequality . . . . . . . . . . . . . . . . . . . 14

1.4.3 Fejér-type inequality . . . . . . . . . . . . . . . . . . . . . . . . 15

II
2 Strongly Reciprocally (p,h)-Convex Function 16

2.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.2 Fundamental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.3 Hermite-Hadamarditypeiinequality . . . . . . . . . . . . . . . . . . . . 23

2.4 Fejér-type inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.5 Fractional integral inequalities . . . . . . . . . . . . . . . . . . . . . . . 29

3 Main Results 39

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.2 The mapping Mw

F (x) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

3.2.1 A New Generalized form of Fejér’s inequality . . . . . . . . . . . 49

3.2.2 AiNew Generalized formiof Hermite-Hadamardiinequality . . . . 60

3.3 The mapping Hw

F (x): . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

4 CONCLUSIONS AND FUTURE RECOMMENDATION 68

III
List of Figures

1.1 ConvexisetivsiNon-convex set . . . . . . . . . . . . . . . . . . . . . . . 4

1.2 Convex function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.3 Epigraphiofitheifunction F . . . . . . . . . . . . . . . . . . . . . . . . . 7

IV
LIST OF SYMBOLS and
ABBREVIATIONS

R The set of real numbers

Rn n-dimensional real space
R+ The set of positive real numbers
I Subset of R
epiF Epigraph of the Function F
F′− Left derivative of the function F
F′+ Right derivative of the function F
SR(ph) Strongly reciprocally (p, h)-convex function
SX(h, I) h-convex function
SV(h, I) h-concave function
Q(I) Godunova–Levin function
K2s s-convex function in the second sense
P(I) p-convexifunction
Mw
F A mapping defined on integrable functions F and w
Hw
F A mapping defined on integrable functions F and w

V
Abstract

This thesis is a detailed study of convex functions, focusing on h-convex, p-

convex, and stronglyireciprocally (p, h)-convex functions of higheriorder cases.iThe
mainiobjective ofithis research isito enhance both theiunderstanding and analysis of
these generalizediconvex functions, with a focus on their connections to established
mathematicaliinequalities, such as theiHermite-Hadamardiinequality, Fejériinequality,
and fractionaliintegraliinequalities. The foundational concept of convexity, its basic
properties, and applications are reexamined. The utility of convex functions in
numerous areas ranging from theoretical research to practical applications of pure
and appliedisciences including, economic models, mathematical analysis, and opti-
mizationiproblems is highlighted. The role of inequality theory is also pointed out
for convex functions. The classical convexity is extended to stronglyireciprocally
(p, h)-convex functions of higheriorder instances, enabling a deeper insight into how
convex functions can be applied to mathematicaliinequalities, suchias theiHermite-
Hadamardiinequality, the Fejériinequality, andifractional integraliinequalities are
explored. A novel mapping Mw
F (x) for h-convex functions is explored along a series

of useful results. These results are formulated through lemmas and propositions,
which are then utilized to derive some generalized Fejér-typeiinequalities and an
improved variant of Hermite-Hadamard inequalities. A second mapping, Hw
F (x) for

1
h-convex function is introduced, with the further contribution of some results that
leads to the derivation of a significant theorem. These findings contribute new tools
for mathematical analysis and significantly broaden the scope of known results to
generalize convex functions and inequalities.

2
Chapter 1

Introductioni

Convexityiplays a key role in mathematical analysis, due to its natural and robust
problem-solving approach in different areas of science such as optimization, economics,
physics, and engineering. Its simple yet rich properties make it an essential component
of study across the broad spectrum of scientific fields.
Before delving into my research work, this chapter includes a general background
of convexity and convex functions, providing readers with a glimpse into modern
approaches to convex analysis, advancement, extensions, and generalizations, illus-
trating how these concepts are applied across different fields. Finally, we discuss some
well-known inequalities for convex functions, such as Hermite-Hadamardiinequality
and Fejér-type inequality.
History contains many concepts that have been essentialiin the developmentiofimathematics.
Ofithese, convexity stands out as one of the exceptional concepts, serving as the
foundation of numerous theories and scientific developments. Its origins can be traced
back to ancient Archimedes and Greece, where Euclidean geometry examined convex
figures and their properties, see [1] and [2] for details. The significance of these

3
was recognized by many famous mathematicians, leading to further exploration and
development. The subsequent mathematicians contributed innumerable extensions
and generalizations of the initially developed concepts, which became solutions to
many problems across multiple science disciplines. The systematic study of convexity
began atitheiend of the nineteenthicentury, with contributionsifrom O. Hölder (1889)
[3], O. Stolz (1893) [4], Ch. Hermite (1883) [5] and J. Hadamard (1893) [6]. J.L.W.V.
Jensen (1905, 1906) [7, 8] was among the first to recognize its value and he made a
structured analysis of convexity and provided its modern definition.
Convexity constitutes two fundamental concepts: convex sets, and convex functions.
Below, I have defined these concepts with some properties.

1.1 Convex Set

Definition 1.1.1. [9] Foriainon-emptyiset I ⊆ Rito be convex, consider any two dis-
tinct points α1 , α2 ∈ I and ζ in [0, 1] we have

α1 ζ + (1 − ζ)α2 ∈ I. (1.1)

Figure 1.1: ConvexisetivsiNon-convex set

4
 In Figure 1.1 we can see the keyidifferences between aiconvex setiandiainon-convex
set. It can be viewed geometrically, every point on the line between any two dis-
tinctipoints initheiset I is also contained initheiset, i.e. ∀α1 , α2 ∈ I and ζ ∈ [0, 1], all
the points on the line between α1 and α2 can be written as z = α1 ζ+(1−ζ)α2 ∈ I, hence
we can say that I contains no indentation or holes. Every real interval canibeiexpressed
asiaiconvexicombination of any two distinctipoints within its interval, that is each real
interval is a convex set. The result of theiintersectionibetween anyicollection of con-
vexisets isiagainiaiconvexiset. Similarly, convexity is preserved between the addition
of two convex sets, and the scalar multiplication of a convex set is just like a vector
space.

1.2 Convex Function

Let’s come to the main idea of a classical convexifunction over a real interval.

Definition 1.2.1. [9] Consider aifunction F : I = [α1 , α2 ] → R, iwhereidomain

̸ I ⊆ R.iTheifunction F isitermed asia
ofithe convexifunction Fiis a convexiseti∅ =
convexifunctioniif itisatisfies theiinequality

(1.2)


F ζa + (1 − ζ)b ≤ ζF(a) + (1 − ζ)F(b), ∀ζ ∈ [0, 1], i

whereia, b∈ [α1 , α2 ] and a ̸= b.

If inequality (1.2) is reversed (i.e. ≤ is changed with ≥), then it leads to the
function being concave. Likewise, if inequality (1.2) ≤ is modified with <, then F
becomes strictlyiconvexifunction.
For aiparticular case, weiobtainiaimidpoint convexifunction for value ζ = 1
2
which is a

5
well-known midpointiconvexifunction as follows:

 
a+b F(a) + F(b)
F ≤ ∀a, b ∈ I.
2 2

Figure 1.2: Convex function

The geometric interpretation of the equation in (1.2) is straightforward. We can

see in Figure 1.2 the line segment joined by two distinct pointsi(a, F(a))iandi(b, F(b))
always lies either on or above the curve ofitheifunction F withinitheiinterval from aitoib.

Example 1.2.1. Below are a few examples of convexifunction F: R → Ri

• Linear Function:iF(x) = ax + b.i

• Quadratic functions:iF(x) = ax2 + bx + c, iwhereia > 0.i

• Power function:iF(x) = xn , ii∀n ≥ 1.i

• Absolute value function:iF(x) = |x|.i

6
 • Exponential function:iF(x) = ebx ,iwhereib isiaiconstant.

Definition 1.2.2 (Epigraph of a function). [9] An epigraph is a set of points positioned

either on or above the graph of theifunctioniF : I ⊆ R → Ridefinedias:

epiF = {(a, t) ∈ I × R : t ≥ F(a)}.

In the following graph, we can visualize the epigraph as a collection of all those
points that belong to the colored region aboveithe graphiof theifunction F.

Figure 1.3: Epigraphiofitheifunction F

1.2.1 Properties of Convex Function

• If theisecondiderivative ofitheifunction Fiis non-negative forievery point

inidomain I then itiisiaiconvexifunction.

7
 • Due to theinature ofitheiconvexifunction, every local minimum within its domain
is guaranteed to be a global minimum.

• A functioniF is known asiconvexiifiand onlyiif the epigraph ofiFiis a convex set.

1.3 Continuity and Differentiability of Convex func-

tions

Continuity and differentiability are essential in the mathematical analysis of con-

vexifunction to understand its behavior and practical use.

1.3.1 Boundedness

Considering a finite convexifunction F on I = [a, b]iisibounded aboveibyiM =

max F(a), F(b) , ianyipointiz ∈ Iicanibe writteniasiz = ζa + (1 − ζ)bi,we obtain the

following relation:

iF(z) ⩽ ζF(a) + (1 − ζ)F(b) ⩽ ζM + (1 − ζ)M = M.i

Moreover,iitihas a lower bound that is determined by expressing anyipointiin

theiform:i a+b
2
+ t.iithen

a + b
1 a + b
1 a+b

iF ⩽ F +t + F − t ,i
2 2 2 2 2

ori
a+b
a + b
a+b

iF + t ⩽ 2F −F − t .i
2 2 2

8
AsiMiis the upperibound, we havei − F[ (a+b)
2
− t] ≥ −Mi, so

a+b
a + b

iF + t ⩽ 2F − M = m.i
2 2

Therefore, convexifunctions are continuousion the interioriof theiridomain, but conti-

nuity at the boundary points is not guaranteed, as the function might display upward
jumps.

1.3.2 Continuity andidifferentiability

The definition of convex functions ensures its smoothness and continuity. Moreover, it
is evident, from the geometrical interpretation and its property of having a non-negative
second derivative. Theifollowing propositionihighlights the geometric significance ofithe
convexifunction:

Proposition 1. [10] ConsideriFibe a convexifunction oni(a, b).iIfia < c < u < d <
b, iand theicorresponding points C = (c, F(c)),iU = (u, F(u)), D = (d, F(d)), i we have
the following relation between the slopes:

F(u) − F(c) F(d) − F(c) F(d) − F(u)

slopeCU = ≤ slopeCD = ≤ slopeDU = .
u−c d−c d−u

Theorem 1.3.1. (see [11]) Consider an open set I =

̸ 0 in R that isia convexiset. Any
convexifunctioniF : I → Rimustibe continuous throughout the domain I.

Theorem 1.3.2. EveryiconvexifunctioniF : I → Riis continuous onian open set I, if

iF′+ (a) ⩽ F′− (b), a ≤ b, i (1.3)

9
where,ithe limit
F(a) − F(a0 )
iF′+ (a) = lim+ , a ∈ I, i
a0 →a a − a0

exist, we say it right derivativeiof Fiatia, isimilarly if limit

F(b) − F(b0 )
iF′− (b) = lim− , b ∈ I, i
b0 →b b − b0

exist, we say it leftiderivative of F at b.

Below is an example ofia discontinuous convexifunction at a specific pointibut con-

tinuous otherwise.

Example 1.3.1. Consider a convexifunction defined as


if u < 0,

0

iF(u) =
u2 if u ≥ 0.i



Another approach to check theicontinuity ofia convexifunction is theiLipschitz con-

dition [10, 12].

Theorem 1.3.3. Consider a convexifunction F : I → RiwhereiIiisia non-emptyiset,

F is considered to be Lipschitz continuous with any real numberil ≥ 0iif itisatisfiesithe
following condition:

i|F(a) − F(b)| ≤ l||a − b|| ∀a, b ∈ I.i

Theidifferentiabilityipropertiesiof convexifunctions depend onithe nature of the

function, some convex functions are differentiable but some are not. Examine deriva-
tives for the best possible results regarding left and right derivatives. Considering a

10
convexifunction F : I → Rihas leftiandiright derivatives at any pointia ∈ I as under

F(a) − F(a0 )
iF′− (a) = lim− ,i (1.4)
a0 →a a − a0

F(a) − F(a0 )
F′+ (a) = lim+ . (1.5)
a0 →a a − a0

Theorem 1.3.4. LetiF : I → Ribe a convexifunction whereiI is aniopeniset, theniF

is increasing ifiandionly ifiF′− (a)iandiF′+ (a)iexist.

Theorem 1.3.5. [1] Consider a convexifunction F : I → R.iThen F isicontinuous on

an open setiIiand it has leftiandiright derivatives on every pointiof I.iAdditionally, for
any two pointsia, b ∈ Iiwe have following relation

iF′− (a) ≤ F′+ (a) ≤ F′− (b) ≤ F′+ (b), i

whereia ≤ b. In particular, bothiF′− andiF′+ are increasing on I.

Here are a few examples of differentiable and non-differentiable convexifunctions.

Example 1.3.2. letiF be a convexifunction

• iF(η) = η 2 isidifferentiable on whole domain.

• iF(η) = eη isidifferentiable on whole domain.

• iF(η) = |η| isinot differentiable atiη = 0.

• iF(η) = max(0, η) is not differentiable at η = 0.

• 
if η < 0, i

0

iF(η) =
if ≥ 0.i

η


11
 The functioniFiis notidifferentiable at η = 0.

Theorem 1.3.6. [10] LetiF : I → Ribeia twice differentiableifunction. The function

F is convexiifiand onlyiif i∀a ∈ I, iF′′ (a) ≥ 0.

1.3.3 Convex functions: Applications, Extensions, and Gener-

alizations

In the above sections, we have explored a detailed analysis of the convexifunction,

highlighting its fundamental properties, continuity, and differentiability details. Due
to these attributes, the convexifunction became pivotal across various applications (see
[1]). Such as in optimization it ensures the local minima as global minima so that it can
be used to model real-world optimization problems, in economics it is used to model
utility and production functions for better efficiency and resource allocation. It aids in
control systems and signal processing in engineering, while in machine learning it helps
to support vector machines, constructing effective loss functions and neural networks.
Most importantly, it has many applications within mathematics such as in calculus
of variations, where it helps to formulate and solve complicated problems that are
required to minimize or maximize specified functionals, it has multiple applications in
geometry and functional analysis, and it contributes in development and refinement of
various mathematical concepts and theorems. Indeed, it contributes useful insights in
many scientific studies due to its adaptability through extensions and generalizations
further enhances its utility across diverse research areas. Some notable generalizations
are p−convexifunctions [13], strongly convexifunctions [14], h−convexifunctions
[15], s−convexifunctions [16], m−convexifunctions [17], log−convexifunctions [18],
η−convexifunctions [19], ϕ−convexifunctions [20], k−convexifunctions [21], many

12
other generalizations have been explored (see [22]), and further developments and
extensions are still underway.

1.4 Some remarkable Inequalities

All the time, inequalities have played a remarkable role in many scientific studies. Very
famous triangular inequality and isoperimetric inequality were discovered by the an-
cient Greeks. In the eighteenth century, eminent scientists such as Gauss, Cauchy,
and Chebyshev provided modern mathematical inequalities formalization and appli-
cations. Over the last few years, inequalities have become powerful problem-solving
tools, widely applied across various fields, including mathematical analysis, physics,
optimization, statistics, and other scientific studies (see [22]). Notable inequalities like
Hölder’siinequality, the powerimeaniinequality, and Jensen’siinequalityiare amongithe
most common and significant in the study of convex functions, which are established
within the subset of real numbers. Several integral inequalities,iincluding the Hermite-
Hadamardiinequality,iFejér-type inequality, Ostrowski inequality, and Gauss inequality,
are formulated for certain functions within convex analysis.

Some fundamental inequalities relatedito theitheory of convexifunctions are subjected

below.

1.4.1 Jensen’siinequality

Jensen’siinequality is among theimost significant inequalitiesiin mathematical analyses

foriconvexifunctions. Thisiwas firstiprovediin 1906 by Johan Ludwig Jensen [23, 8].

13
Consider aiconvexifunctioniF : I ⊆ R → R, iJensen’s inequality states that:

m
X  Xm
iF pi ai ≤ pi F(ai ), i (1.6)
i=1 i=1

where ai ∈ I, ipi ≥ 0iwithi pi = 1iandii = 1, . . . , m. This inequality shows the

Pm
i=1

behavior of the convexifunction, the function evaluates at the weighted average is al-
ways less or equalito the weighted averageiof theifunction value on every point ofithe
interval.
It has become an inspiration for many mathematicians and scientists. They ex-
tended this inequality in the form of different generalizations and inequalities i.e. re-
verse Jensen’siinequality, Jensen’siinequality for Expectation with Convex functions,
Hölder’siinequality, and Minkowski’siinequality. It has many applicationsiin numerous
fieldsisuch as statistics, mathematics, physics, optimization, and economics.

1.4.2 Hermite-Hadamardiinequality

The classicaliHermite-Hadamardiinequality was establishedibyiHermite in 1881 [5],

andiLater, ini1893 [6] Hadamardiexplored the same inequality, yet; he was not familiar
with Hermite’s result. LetiF : I ⊆ R → R, iI =
̸ ∅ibe a convexifunction, for a, b ∈ I,
theiinequalityiholds

  Z b
a+b F(a) + F(b)
iF ≤ F(x)dx ≤ .i (1.7)
2 a 2

Hermite-Hadamardiinequality estimates, the integral average over a compact interval

of a convexifunction and establishes upper andilower bounds forithe function’sivalue
at the midpoint andiendpoint of the interval. Due to these valuable properties, many

14
researchers have made significant results and generalizations in different domains of
mathematics. These generalizations helped to investigate many problems developed in
applied mathematics, analyses, optimization, and economics. Given below is a weighted
version of Hermite-Hadamardiinequality.

1.4.3 Fejér-type inequality

In 1906, Lipót Fejér introduced a generalized version of Hermite-Hadamardiinequality

while studying trigonometric polynomials. This famous integral inequality is known as
Fejér’s inequalities [24], where a convexifunction along the concept of non-negative sym-
metric weights is involved. Consideria convexifunctioniF : [α1 , α2 ] = I ⊆ R → Riand
weight w : [α1 , α2 ] → R, Fejér’s inequality states that:

Z b Z b
F(a) + F(b) b
 Z
a+b
iF w(x)dx ≤ F(x)w(x)dx ≤ w(x)dx, i (1.8)
2 a a 2 a

Rb
where,iwiisinon-negative, integrable weightisuch thati a
w(x)dx > 0, i also it is sym-
metriciwith respectito a+b
2
implies that, wisatisfiesiw(x) = w(a + b − x)ii∀x ∈ [α1 , α2 ].
Foriw(x) = 1 it becomes Hermite-Hadamardiinequality in equation (1.7).
This contribution had a lasting impact on mathematical analysis. It has numerous
applications and generalizations in many fields i.e. numerical methods, approximation
theory, statistics, discrete mathematics, and beyond, see [25, 26, 27, 28, 29, 30, 31, 32,
33, 34, 35] for more detail.

15
Chapter 2

Strongly Reciprocally (p,h)-Convex

Function

2.1 Preliminaries

Over the past century, the conceptiof convexifunction has beeniextended inivarious
dimensions, either by modifying the originally established inequalities or generaliz-
ing them to abstract spaces. To establish a strong foundation for the later chapters,
this section introduces fundamental definitions and preliminary concepts related to
stronglyireciprocallyi(p, h)-convexifunctions.

Definition 2.1.1 (p-convex set [13]). Aiset I = [α1 , β1 ] ⊆ R \ {0}iis aip-convex set if

1p
ζap + (1 − ζ)bp ∈ I, i (2.1)

foriall a, b ∈ I and ζ ∈ [0, 1],iwhereip = 2u + 1ior p = dc , d = 2v + 1, ic = 2w + 1, i

andiu, v, w ∈ N.i

16
Definition 2.1.2 (p-convexifunction [13]). LetiI = [α1 , β1 ] ⊆ R \ {0}i be a
p-convexiset.iA functioniF : I = [α1 , β1 ] ⊆ R isip-convex if

1p

iF ζap + (1 − ζ)bp ≤ ζF(a) + (1 − ζ)F(b), i (2.2)

foriallia, b ∈ I and ζ ∈ [0, 1].

Definition 2.1.3 (Harmonic convex set [36]). A set I = [α1 , β1 ] ⊆ R\{0} is harmonic
convex if
ab
i ∈ I, i (2.3)
ζa + (1 − ζ)b

foriall a, b ∈ I andiζ ∈ [0, 1].

Definition 2.1.4 (Harmonic convex function [36]). Let I = [α1 , β1 ] ⊆ R \ {0} beia
harmonic convexiset. AifunctioniF : I = [α1 , β1 ] ⊆ Ri is harmoniciconvex if

ab
(2.4)


iF ≤ (1 − ζ)F(a) + ζF(b), i
ζa + (1 − ζ)b

foriallia, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.5 (Strongly convex function [37]). AifunctioniF : I = [α1 , β1 ] ⊆ Riis

a stronglyiconvexifunction with modulusiχ ≥ 0ion I if

iF ζa + (1 − ζ)b ≤ ζF(a) + (1 − ζ)F(b) − χζ(1 − ζ)(b − a)2 , i (2.5)

foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.6 (Strongly p-convex function [38]). AifunctioniF : I = [α1 , β1 ] ⊆ Riis

17
stronglyip-convexiif

1p

iF ζap + (1 − ζ)bp ≤ ζF(a) + (1 − ζ)F(b) − χζ(1 − ζ)(bp − ap )2 , i (2.6)

foriallia, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.7 (p-Harmonic convex set [39]). A set I = [α1 , β1 ] ⊆ R \ {0} is

aip-harmoniciconvex set if

 1p
ap bp

i ∈ I, i (2.7)
ζap + (1 − ζ)bp

foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.8 (p-Harmonic convex function [39]). Let I = [α1 , β1 ] ⊆ R \ {0} be

aip-harmoniciconvexiset. Aifunction F : I = [α1 , β1 ] ⊆ Riis p-harmoniciconvex if

 1p 
ap bp

iF ≤ (1 − ζ)F(a) + ζF(b), i (2.8)
ζap + (1 − ζ)bp

foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.9 (Strongly reciprocally convex function [40]). LetiI be aniinterval,

and let χ ∈ (0, ∞). A functioniF : I = [α1 , β1 ] ⊆ Ri is stronglyireciprocally convexiwith
modulus χ on I if

   2
ab 1 1
iF ⩽ (1 − ζ)F(a) + ζF(b) − χζ(1 − ζ) − ,i (2.9)
ζa + (1 − ζ)b a b

foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.10 (Stronglyireciprocally p-convexifunction [41]). Let I beia p-

convexiset, andiletiχ ∈ (0, ∞). Aifunction F : I = [α1 , β1 ] ⊆ R is stronglyireciprocally

18
p-convexifunction withimodulus χ on Iiif

 1p  2
ap bp
 
1 1
iF ⩽ (1 − ζ)F(a) + ζF(b) − χζ(1 − ζ) p − p ,i (2.10)
ζap + (1 − ζ)bp a b

foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.11 (h-convex function [15]). Let I, h : ζ = [α1 , β1 ] → R be non-negative

functions.iTheniF isih-convexiif

(2.11)


iF ζa + (1 − ζ)b ≤ h(ζ)F(a) + h(1 − ζ)F(b), i

foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.12 ((p, h)-convexifunction [42]). Let I = [α1 , β1 ] ⊆ R \ {0}ibe a p-

convexiset. A functioniF : I = [α1 , β1 ] ⊆ Riis (p, h)-convexifunctioniif Fiis non-
negativeiand

1p

iF ζap + (1 − ζ)bp ≤ h(ζ)F(a) + h(1 − ζ)F(b), i (2.12)

foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.13 (higher-order Strongly convex [43]). LetiI beian interval,iand let
χ ∈ (0, ∞).iAifunctioniF : I = [α1 , β1 ] ⊆ R is higher-order stronglyiconvexiwith mod-
ulus χ on I ifi

iF ζa + (1 − ζ)b ≤ ζF(a) + (1 − ζ)F(b) − χϕ1 (ζ)∥a − b∥l , i (2.13)

foriall a, b ∈ Iiandiζ ∈ [0, 1] and l ≥ 1,iwhere ϕ1 (ζ) = ζ(1 − ζ).i

Definition 2.1.14 (Strongly reciprocally (p, h)-convexifunction [44]). LetiIibe

aniinterval, and let χ ∈ (0, ∞).iA functioniF : I = [α1 , β1 ] ⊆ Riis stronglyireciprocally

19
(p, h)-convexiwith modulusiχ on I if i

 1p  2
ap bp
 
1 1
iF ⩽ h(1 − ζ)F(a) + h(ζ)F(b) − χζ(1 − ζ) p − p , i
ζap + (1 − ζ)bp a b
(2.14)
foriall a, b ∈ Iiand ζ ∈ [0, 1].i

Definition 2.1.15 (Stronglyireciprocally (p, h)-convexifunctioniofihigher-order

[44]). Let I beian interval,iand letiχ ∈ (0, ∞).iAifunction F : I = [α1 , β1 ] ⊆ Riis
stronglyireciprocallyi(p, h)-convexiof higher-orderiwith modulusiχion I ifi

 1p  l
ap bp

1 1
iF ⩽ h(1 − ζ)F(a) + h(ζ)F(b) − χϕ1 (ζ) p − p , i (2.15)
ζap + (1 − ζ)bp a b

foriall a, b ∈ I, ζ ∈ [0, 1],iand l ≥ 1,iwhere ϕ1 (ζ) = ζ(1 − ζ).i

Remark 1. Insertingil = 2 intoiDefinition 2.1.15 withiϕ1 (ζ) asiabove, weiobtain Def-

initioni2.1.14.iSimilarly,iinsertingil = 2iandih(ζ) = ζiintoiDefinition 2.1.15,weiobtain
Definitioni2.1.10, andil = 2, ih(ζ) = ζiandip = 1, iDefinition 2.1.15, reducesitoiDefinition
2.1.9.
Asiwe know, R isia normedispace withithe standardimodulus normiapplied. Thusifor
anyia ∈ R,

∥a∥ = |a|. (2.16)

Using (2.16), inequality 2.1.15 can be expressed as

 1p  l
ap bp

1 1
iF ⩽ h(1 − ζ)F(a) + h(ζ)F(b) − χϕ1 (ζ) − ,i (2.17)
ζap + (1 − ζ)bp ap bp

foriallia, b ∈ I, ζ ∈ [0, 1], iand l ≥ 1, where ϕ1 (ζ) = ζ(1 − ζ).

20
2.2 Fundamental Results

Onward in this section, we explore an interesting domain called stronglyireciprocally

(p, h)-convexifunctions of higheriorder, which isinamed as SR(ph) for the rest of
this thesis. We obtained some useful results by performingisimple algebraic opera-
tionsioniSR(ph).
Theiaddition of two functions from SR(ph) is obtained in the following proposition.

Proposition 2. [44] Consider two function from SR(ph), F,G : I → R withimodulus

χion I. TheniF,G : I → Riis alsoiin SR(ph)iwithimodulus χ∗ on I, whereiχ∗ = 2χi.

Proof. Weibegin by the definitioniprovided:

 1p   1p   1p 
ap bp ap bp ap bp
  
F+G =F +G
ζap + (1 − ζ)bp ζap + (1 − ζ)bp ζap + (1 − ζ)bp
l
1 1
≤ h(ζ)F(a) + h(1 − ζ)F(b) − χϕ1 (ζ) p − p
a b
l
1 1
+ h(ζ)G(a) + h(1 − ζ)G(b) − χϕ1 (ζ) p − p ,
a b
(2.18)
which is further simplified as

l
1 1
= h(ζ)(F+G)(a) + h(1 − ζ)(F + G)(b) − 2χϕ1 (ζ) p − p
a b
l
1 1
= h(ζ)(F + G)(a) + h(1 − ζ)(F + G)(b) − χ∗ ϕ1 (ζ) − ,
ap bp

whereiχ∗ = 2χ, χ ≥ 0 and ϕ1 (ζ) = ζ(1 − ζ).iHence theiproof is established.

The result for scalar multiplication in SR(ph) is achieved in following proposition.

21
Proposition 3. [44] Considering aifunctioniiniSR(ph)iF : I → Riwithimodulusiχ ≥
0. Theniforiany λ ≥ 0,iλF : I → RiisialsoiiniSR(ph) withimodulus ψ ∗ ion I,
whereiψ ∗ = λχ.

Proof. Consideriλ ≥ 0.iTherefore, byigiven definitioniof F weihave

 1p   1p 
ap bp ap bp
  
λF =λ F
ζap + (1 − ζ)bp ζap + (1 − ζ)bp
 l
1 1
≤ λ h(ζ)F(a) + h(1 − ζ)F(b) − χϕ1 (ζ) p − p
b a
l
1 1
= h(ζ)λF(a) + h(1 − ζ)λF(b) − λχϕ1 (ζ) −
bp ap
l
1 1
= h(ζ)λF(a) + h(1 − ζ)λF(b) − ψ ∗ ϕ1 (ζ) p
− p ,
b a

whereiψ ∗ = λχ, χ ≥ 0, iandiϕ1 (ζ) = ζ(1 − ζ). Hence theiproof is established.

Proposition 4. [44] Considering Fi : I → R, 1 ≤ i ≤ n, ibeiiniSR(ph)

withimodulusiχ. Theniforiλi ≥ 0, i1 ≤ i ≤ n, itheifunction F : I → R, iwhereiF =
Pn Pn
i=1 λ i Fi, iisialsoiiniSR(ph)iwithimodulusiγ ≥ 0,iwhereiγ = i=1 λi χ.

Proof. LetiI beia p-harmoniciconvexiset. Thenifor all a, b ∈ I ans ζ ∈ [0, 1], weihave

 1p  n  1p 
ap bp ap bp
 X 
F = λi Fi
ζap + (1 − ζ)bp i=1
ζap + (1 − ζ)bp
n  l
X 1 1
≤ λi h(ζ)Fi (a) + h(1 − ζ)Fi (b) − χϕ1 (ζ) p − p
i=1
b a
n n n  l
X X X 1 1
= h(ζ) λi Fi (a) + h(1 − ζ) λi Fi (b) − λi χϕ1 (ζ) p − p
i=1 i=1 i=1
b a
l
1 1
= h(ζ)F(a) + h(1 − ζ)F(b) − γϕ1 (ζ) p
− p ,
b a

where γ = λi χ. Hence this is the required result.

Pn
i=1

22
Proposition 5. [44] Considering Fi : I → R,i1 ≤ i ≤ nibeiiniSR(ph) withimodulusiχ.
TheniF = max{Fi , i = 1, 2, . . . , n}iisialsoiin SR(ph)iwithimodulusiχ.

Proof. LetiI beia p-harmoniciconvexiset. Thenifor all a, b ∈ Iiand ζ ∈ [0, 1], weihave

 1p 
ap bp

F
ζap + (1 − ζ)bp
 1p 
ap bp
  
= max Fi , i = 1, 2, 3, . . . , n
ζap + (1 − ζ)bp
ap bp
 
= Fc
ζap + (1 − ζ)bp
l
(2.19)
1 1
≤ h(ζ)Fc (a) + h(1 − ζ)Fc (b) − χϕ1 (ζ) p − p
b a
l
  1 1
= h(ζ) max Fi (a) + h(1 − ζ) max Fi (b) − χϕ1 (ζ) p − p
b a
l
1 1
= h(ζ)F(a) + h(1 − ζ)F(b) − χϕ1 (ζ) − .
bp ap

Hence the proof is established.

2.3 Hermite-Hadamarditypeiinequality

Inithis section,iwe have proved a generalized form ofitheiHermite-Hadamarditypeiforithe

function SR(ph).

Theorem 2.3.1. [44] LetiI ⊂ R \ {0}ibeian interval.iIf F : I → RibeiiniSR(ph)

withimodulusiχ ≥ 0iandiF ∈ L[x, y],ithen for h( 21 ) ̸= 0, weihave

23
 1 l
2xp y p p 1 y − xp 1 − (−1)2l+1
    p 
1
F p + χϕ1
2h( 12 ) x + yp 2 xp y p 2(l + 1)
y
p(xp y p )
Z
F(a)
≤ p da (2.20)
y − x x a1+p
p
Z 1 lZ 1
  y p − xp
≤ h(1 − ζ)F(x) + h(ζ)F(y) dζ − χ ϕ1 (ζ) dζ.
0 xp y p 0

Proof. Substituting ζ = 1
2
in the Definition 2.1.15. We have

 p p  1p        l
2a b 1 1 1 1 1
F ⩽ h F(a) + h F(b) − χϕ 1 − . (2.21)
ap + bp 2 2 2 ap bp

p p 1 p p 1
Let a = [( ζxp +(1−ζ)y
x y
p ) ] and b = [( ζy p +(1−ζ)xp ) ]. Integrating 2.21 with respect to ζ over
p x y p

[0, 1], weihave

1  1p   1p 
2xp y p p xp y p xp y p
      
1 1
F p ≤h F +h F
x + yp 2 ζxp + (1 − ζ)y p 2 ζxp + (1 − ζ)xp
  p l
1 y − xp
− χϕ1 p p
|1 − 2ζ|l ,
2 xy
Z 1  1  1p 
1
2xp y p p xp y p
 Z   
1
F p dζ ≤ h F dζ
0 x + yp 0 2 ζxp + (1 − ζ)y p
Z 1     1p 
1 xp y p
+ h F dζ
0 2 ζy p + (1 − ζ)xp
  p lZ 1
1 y − xp
− χϕ1 |1 − 2ζ|l dζ,
2 xp y p 0
 1p p p Z y l
p p
1 y − xp 1 − (−1)2l+1
     p 
2x y 1 p(x y ) F(a)
F p ≤ 2h da − χϕ1 ,
x + yp 2 y p − xp x a1+p 2 xp y p 2(l + 1)

and

24
 1 l
2xp y p p 1 y − xp 1 − (−1)2l+1
   p 
F p + χϕ1
x + yp 2 xp y p 2(l + 1)
  p p Z y
1 p(x y ) F(a)
≤ 2h da,
2 y p − xp x a1+p
1 l
2xp y p p 1 y − xp 1 − (−1)2l+1
    p 
1
F p + χϕ1
2h( 12 ) x + yp 2 xp y p 2(l + 1)
p p Z y
p(x y ) F(a)
≤ p da,
y − x x a1+p
p

the required leftiside ofiinequality (2.20)iis established.

Furthermore, inithe rightiside ofiinequality (2.20), settingia = xiand b =
yiiniDefinition 2.1.15 givesi

 1p  l
xp y p

1 1
iF ⩽ h(1 − ζ)F(x) + h(ζ)F(y) − χϕ1 (ζ) − p . (2.22)
txp + (1 − t)y p x p y

Integrating (2.22) withirespect toiζ over [0,1], weiget

1  1p  1 1
xp y p
Z  Z Z
F dζ ≤ h(1 − ζ)F(x) dζ + h(ζ)F(y) dζ
0 txp + (1 − t)y p 0 0
l 1
y p − xp
Z
−χ ϕ1 (ζ) dζ
xp y p 0

and

y 1 l 1
p(xp y p ) y p − xp
Z Z Z
F(a)
da ≤ h(1 − ζ)F(x) + h(ζ)F(y) dζ − χ ϕ1 (ζ) dζ,
y p − xp x a1+p 0 xp y p 0

the required right side of inequality (2.20) is established here. Thus, the proof is now
complete.

Remark 2. 1. Settingil = 2,ih(ζ) = ζ, andip = 1,iwith ϕ1 (ζ) = ζ(1 − ζ)

intoiTheorem 2.3.1, WeigetitheiHermite–HadamardiinequalityiforiStronglyireciprocally

25
 convexifunction; (see [40], Theorem 3.1)

2. Setting l = 2, h(ζ) = ζ, p = 1,iand χ = 0iwithiϕ1 (ζ) = ζ(1 − ζ) intoiTheorem

2.3.1 weiobtain the Hermite-Hadamardiinequalityifor harmonic convexifunctions;
(see [40], Theorem 2.4).

2.4 Fejér-type inequality

Now, forithe functionsibelonging toiSR(ph) we will develop Fejér-type inequality.

Theorem 2.4.1. [44] Considering I ⊂ R \ {0} beian interval.iIf F : I → R is in

SR(ph) withimodulusiχ ≥ 0, thenifor h( 12 ) ̸= 0, weihave

1 Z
2xp y p p y w(a)
 
1
F p da
2h( 12 ) x + yp x a
1+p
 Z y
|2xp y p − (xp + y p )ap |l w(a)

χ 1
+ p p l ϕ1 da
|x y | 2 x |ap |l a1+p
Z y
F(a)w(a)
≤ da (2.23)
x a1+p
 y
 p p
x (y − ap ) w(a)
Z 

≤ F(x) + F(y) h p p da
x x (y − xp ) a1+p
lZ y
y p − xp
 p p
x (y − ap ) w(a)

−χ ϕ1 p p da
xp y p x a (y − xp ) a1+p

forix, y ∈ Iiwith x ≤ y and F ∈ L[x, y],iwhere w : I → R isia non-negativeiintegrable

functionisatisfying
 1p  1p 
xp y p xp y p
 
w =w .
ap xp + y p − a p
Proof. Substituting ζ = 1
2
into Definition 2.1.15 yields

 p p  1p        l
2a b 1 1 1 1 1
F ⩽ h F(a) + h F(b) − χϕ 1 − . (2.24)
ap + bp 2 2 2 ap bp

26
 p p 1 p p 1
Let a = [( ζxp +(1−ζ)y
x y
p ) ] and b = [( ζy p +(1−ζ)xp ) ]. Integrating (2.24) with respect to ζ
p x y p

over [0, 1], we have

1  1p   1p 
2xp y p p xp y p xp y p
      
1 1
F p ≤h F +h F
x + yp 2 ζxp + (1 − ζ)y p 2 ζy p + (1 − ζ)xp
  p l
1 ζx + (1 − ζ)y p ζy p + (1 − ζ)xp
− χϕ1 − .
2 xp y p xp y p

Sinceiw isia non-negativeisymmetric andiintegrableifunction, weihave

 1   1p 
2xp y p p xp y p

F p w
x + yp ζxp + (1 − ζ)y p
 1p    1p 
xp y p xp y p
  
1
≤h F w
2 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
1  1p 
xp y p xp y p
    
1 p
+h F w
2 ζy p + (1 − ζ)xp ζxp + (1 − ζ)y p
  p l   1p 
1 ζx + (1 − ζ)y p ζy p + (1 − ζ)xp xp y p
− χϕ1 − w .
2 xp y p xp y p ζxp + (1 − ζ)y p
(2.25)
Integrating inequality (2.25) withirespectito ζ over [0, 1], weihave

1  1   1p 
2xp y p p xp y p
Z 
F p w dζ
0 x + yp ζxp + (1 − ζ)y p
Z 1     1p    1p 
1 xp y p xp y p
≤ h F w dζ
0 2 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1     1p    1p 
1 xp y p xp y p
+ h F w dζ
0 2 ζy p + (1 − ζ)xp ζxp + (1 − ζ)y p
 Z 1 l   1p 
1 ζxp + (1 − ζ)y p ζy p + (1 − ζ)xp xp y p
− χϕ1 − w dζ,
2 0 xp y p xp y p ζxp + (1 − ζ)y p

and

27
 1 Z
2xp y p p y w(a)
 Z y
|2xp y p − (xp + y p )ap |l w(a)

χ 1
F p 1+p
da + ϕ1 da
x + yp x a |xp y p |l 2 x |ap |l a1+p
 Z y
1 F(a)w(a)
≤ 2h da,
2 x a1+p
1 Z
2xp y p p y w(a)
 
1
F p da
2h( 12 ) x + yp x a
1+p
 Z y
|2xp y p − (xp + y p )ap |l w(a)

χ 1
+ p p l ϕ1 da
|x y | 2 x |ap |l a1+p
Z y
F(a)w(a)
≤ da,
x a1+p

this is the required left side if inequality (2.23).

Further, for the right side of inequality (2.23), let a = x in Definition 2.1.15 we get

 1p  l
xp y p

1 1
F ⩽ h(1 − ζ)F(x) + h(ζ)F(y) − χϕ1 (ζ) − p .
ζxp + (1 − ζ)y p x p y

Since w isia noninegative symmetriciand integrableifunction, we have

 1p    1p 
xp y p xp y p

F w
ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
 1p   1p 
xp y p xp y p
 
⩽ h(1 − ζ)F(x)w + h(ζ)F(y)w
ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
l   1p 
1 1 xp y p
− χϕ1 (ζ) p − p w .
x y ζxp + (1 − ζ)y p
(2.26)
Integrating inequality (2.26) withirespect toiζ over [0, 1], weiobtain

1  1p    1p 
xp y p xp y p
Z 
F w dζ
0 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1  1p 
xp y p

⩽ h(1 − ζ)F(x)w dζ
0 ζxp + (1 − ζ)y p

28
 1  1p 
xp y p
Z 
+ h(ζ)F(y)w dζ
0 ζxp + (1 − ζ)y p
Z 1 l   1p 
1 1 xp y p
−χ ϕ1 (ζ) p − p w dζ,
0 x y ζxp + (1 − ζ)y p
and y y  p p
x (y − ap ) w(a)
Z Z 
F(a)w(a)

da ≤ F(x) + F(y) h p p da
x a1+p x a (y − xp ) a1+p
lZ y
y p − xp
 p p
x (y − ap ) w(a)

−χ ϕ1 p p da,
xp y p x a (y − xp ) a1+p
this isithe required rightiside ofiinequality (2.23). Thisicompletes theiproof.

Remark 3. (i) Settingil = 2, h(ζ) = ζ, andip = 1iwith ϕ1 (ζ) = ζ(1 − ζ)

intoiTheorem 2.4.1, we get Fejér-typeiinequalityifor Stronglyireciprocally
convexifunctions; (seei[40], Theorem 3.7).

(ii) Inithe sameifashion theiinsertion of l = 2 andih(ζ) = ζ with ϕ1 (ζ) = ζ(1 −

ζ) intoiTheorem 2.4.1 producesiFejér-type inequalityiforiStronglyireciprocallyip-
convexifunctions;i(see [41], Theorem 3.5).

2.5 Fractional integral inequalities

In the past years, Fractional integral inequalities have gained significant popularity
and importance in many fields of science and engineering, to study see [29, 45, 46, 47].
Now we establish someifractional integraliinequalities forifunctions withiderivatives
iniSR(ph). Toiobtain resultsiof ouridesired type, weineed theifollowing lemma,iwhich
canibe foundiin [48] Lemma 2.1.

Lemma 2.5.1. [48] Let F : I = [α1 , β1 ] ⊆ R beia differentiableifunctionion theiinterior

29
I o iof I.iIf F′ ∈ L[x, y]iand λ ∈ [0, 1], theni

1 
2xp y p p p(xp y p ) y F(a)
   Z
F(x + F(y)
(1 − λ)F +λ − p da
xp + y p 2 y − xp x a1+p
Z 1 1+ 1p   1p 
(y p − xp ) xp y p xp y p

2
′
= (2ζ − λ) F dζ
2p(xp y p ) 0 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1 1+ 1p   1p  
xp y p xp y p

′
+ (2ζ − 2 + λ) p + (1 − ζ)y p
F p + (1 − ζ)y p
dζ .
1
2
ζx ζx
(2.27)

Theorem 2.5.2. [44] Let I = [α1 , β1 ] ⊆ R\{0} beia p-harmoniciconvexiset, andilet

F : I → Ribe aidifferentiable functionion theiinterior I o ofiI.iIf F′ ∈ L[x, y]iand
|F′ |q is aiStrongly reciprocallyi(p, h)-convexifunctioniofihigher orderion I, q ≥ 1,iand
λ ∈ [0, 1], theni

1 
2xp y p p p(xp y p ) y F(a)
   Z
F(x) + F(y)
(1 − λ)F +λ − p da
xp + y p 2 y − xp x a1+p
(y p − xp )  1 q q
≤ C1 (p, x, y)1− q C3 (p, x, y) F′ (x) + C5 (p, x, y) F′ (y)
p
2p(x y )p
(2.28)
1 1 q q
+ C7 (p, x, y)χ q + C2 (p, y, x)1− q C6 (p, y, x) F′ (x) + C4 (p, y, x) F′ (y)
1 
+ C8 (p, y, x)χ q ,

where

1 1+ 1p
xp y p
Z 
2
C1 (p, x, y) = |2ζ − λ| dζ, (2.29)
0 ζxp + (1 − ζ)y p
Z 1 1+ 1p
xp y p

C2 (p, y, x) = |2ζ − 2 + λ| dζ, (2.30)
1
2
ζxp + (1 − ζ)y p
Z 1 1+ 1p
xp y p

2
C3 (p, x, y) = h(1 − ζ)|2ζ − λ| dζ, (2.31)
0 ζxp + (1 − ζ)y p

30
 1 1+ 1p
xp y p
Z 
C4 (p, y, x) = h(ζ)|2ζ − 2 + λ| p + (1 − ζ)y p
dζ, (2.32)
1
2
ζx
Z 1 1+ 1p
xp y p

2
C5 (p, x, y) = h(ζ)|2ζ − λ| dζ, (2.33)
0 ζxp + (1 − ζ)y p
Z 1 1+ 1p
xp y p

C6 (p, y, x) = h(1 − ζ)|2ζ − 2 + λ| p + (1 − ζ)y p
dζ, (2.34)
1
2
ζx
Z 1 1+ 1p l
xp y p

2 1 1
C7 (p, x, y) = − ϕ1 (ζ)|2ζ − λ| − dζ, (2.35)
0 ζxp + (1 − ζ)y p y p xp
Z 1 1+ 1p l
xp y p

1 1
C8 (p, y, x) = − ϕ1 (ζ)|2ζ − 2 + λ| p + (1 − ζ)y p p
− p
dζ. (2.36)
1
2
ζx y x

Proof. Applying Lemma 2.5.1, weihave

1 
2xp y p p p(xp y p ) y F(a)
   Z
F(x) + F(y)
(1 − λ)F +λ − p da
xp + y p 2 y − xp x a1+p
Z 1 1+ 1p  1p 
(y p − xp ) xp y p xp y p
 
2
′
≤ (2ζ − λ) F dζ
2p(xp y p ) 0 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1 1+ 1p  1p 
xp y p xp y p
  
′
+ (2ζ − 2 + λ) p + (1 − ζ)y p
F p + (1 − ζ)y p
dζ .
1
2
ζx ζx

Now applying power mean inequality,

Z 1 1+ 1p 1− 1q

(y p − xp ) xp y p

2
≤ (2ζ − λ) dζ
2p(xp y p ) 0 ζxp + (1 − ζ)y p
Z 1 1+ 1p   1p  q  1q
xp y p xp y p

2
′
× (2ζ − λ) F dζ
0 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1 1+ 1p 1− 1q
xp y p

+ (2ζ − 2 + λ) dζ
1
2
ζxp + (1 − ζ)y p
Z 1 1+ 1p   1p  q  1q 
xp y p xp y p

′
× (2ζ − 2 + λ) p + (1 − ζ)y p
F p + (1 − ζ)y p
dζ .
1
2
ζx ζx

31
Since |F′ (a)|q isiin SR(ph), nowiwe’ve

Z 1 1+ 1p 1− 1q

(y p − xp ) xp y p

2
≤ (2ζ − λ) dζ
2p(xp y p ) 0 ζxp + (1 − ζ)y p
Z 1 1+ 1p 
xp y p

2 q q
× (2ζ − λ) p + (1 − ζ)y p
h(1 − ζ) F′ (x) + h(ζ) F(y)
0 ζx
l  1q Z 1 1+ 1p 1− 1q
xp y p

1 1
− χϕ1 (ζ) p − p dζ + (2ζ − 2 + λ) p + (1 − ζ)y p
dζ
y x 1
2
ζx
Z 1  p p
1+ 1p 
xy q q
× (2ζ − 2 + λ) h(1 − ζ) F′ (x) + h(ζ) F′ (b1 )
1
2
ζx + (1 − ζ)y
p p

l  1q 
1 1
− χϕ1 (ζ) p − p dζ
y x
(y p − xp )  1 q q
= p p
C1 (p, x, y)1− q C3 (p, x, y) F′ (x) + C5 (p, x, y) F′ (y)
2p(x y )
1 1 q q
+ C7 (p, x, y)χ q + C2 (p, y, x)1− q C6 (p, y, x) F′ (x) + C4 (p, y, x) F′ (y)
1 
+ C8 (p, y, x)χ q ,

hence, this is required result.

Remark 4. In Theorem 2.5.2, inserting h(ζ) = ζ, χ = 0, and l = 2 with ϕ1 (ζ) =

ζ(1 − ζ), we obtain Theorem 2.2 in reference [48].

Now substituting q = 1,iTheorem 2.5.2 reducesito theifollowing result.

Corollary 2.5.3. [44] iLet I = [α1 , β1 ] ⊆ R\{0} beia p-harmoniciconvexiset, andilet

F : I = [α1 , β1 ] ⊆ R\{0} → R beia differentiableifunction onithe interioriI o iofiI.
IfiF′ ∈ L[x, y], i|F′ |q iis iniSR(ph)ion I, andiλ ∈ [0, 1], theni

32
 1 
2xp y p p p(xp y p ) y F(x)
   Z
F(x) + F(y)
(1 − λ)F +λ − p dx
xp + y p 2 y − xp x x1+p
(y p − xp ) 
′ (2.37)
≤ C 3 (p, x, y) + C 6 (p, y, x) F (x)
2p(xp y p )
+ C5 (p, y, x) + C4 (p, x, y) F′ (y) + C7 (p, x, y) + C8 (p, y, x) χ ,




where C3 , C4 , C5 , C6 , C7 and C8 are given by (2.31)-(2.36).

Remark 5. In Corollary 2.5.3 inserting h(ζ) = ζ, χ = 0, iandil = 2 withiϕ1 (ζ) =

ζ(1 − ζ), we obtainiCorollary 2.3 in reference [48].

Theorem 2.5.4. [44] LetiI = [α1 , β1 ] ⊆ R\{0}ibeia p-harmoniciconvexiset, andilet

F : I = [α1 , β1 ] ⊆ R\{0} → Ribeia differentiableifunction onithe interior I o iofiI.iIf
F′ ∈ L[x, y] isiStrongly reciprocallyi(p, h)-convexifunction ofihigheriorderion I, r, q >
1,i 1r + 1
q
= 1, andiλ ∈ [0, 1], itheni

1 
2xp y p p p(xp y p ) y F(a)
   Z
F(x) + F(y)
(1 − λ)F +λ − p da
xp + y p 2 y − xp x a1+p
1
(y p − xp )
 r+1
λ + (1 − λ)r+1 r  q
≤ × C9 (q, p; x, y) F′ (x)
2p(xp y p ) 2(r + 1)
q
1 (2.38)
+ C11 (q, p; x, y) F′ (y) + C13 (q, p; x, y)χ q
q q
+ C12 (q, p; y, x) F′ (x) + C10 (q, p; y, x) F′ (y)

1 
+ C14 (q, p; y, x)χ q ,

where

1 q+ pq
xp y p
Z 
2
C9 (q, p; x, y) = h(1 − ζ) dζ, (2.39)
0 ζxp + (1 − ζ)y p
Z 1 q+ pq
xp y p

C10 (q, p; y, x) = h(ζ) dζ, (2.40)
1
2
ζxp + (1 − ζ)y p

33
 1 q+ pq
xp y p
Z 
2
C11 (q, p; x, y) = h(ζ) dζ, (2.41)
0 ζxp + (1 − ζ)y p
Z 1 q+ pq
xp y p

C12 (q, p; y, x) = h(1 − ζ) p + (1 − ζ)y p
dζ, (2.42)
1
2
ζx
Z 1 q+ pq l
xp y p

2 1 1
C13 (q, p; x, y) = − ϕ1 (ζ) − dζ, (2.43)
0 ζxp + (1 − ζ)y p y p xp
Z 1 q+ pq l
xp y p

1 1
C14 (q, p; y, x) = − ϕ1 (ζ)|2ζ − 2 + λ| − p dζ. (2.44)
1
2
ζx + (1 − ζ)y
p p y p x

Proof. Using Lemma 2.5.1, weihave

 1p
2xp y p p(xp y p ) y F(a)
   Z
F(x) + F(y)
(1 − λ)F +λ − p da
xp + y p 2 y − xp x a1+p
Z 1 1+ 1p  1p 
(y p − xp ) xp y p xp y p
 
2
′
≤ (2ζ − λ) F dζ
2p(xp y p ) 0 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1 1+ 1p  1p 
xp y p xp y p
  
′
+ (2ζ − 2 + λ) p + (1 − ζ)y p
F p + (1 − ζ)y p
dζ .
1
2
ζx ζx

UsingiHölder’s integral inequality, we get

Z 1  1r
(y p − xp ) 2 r
≤ (2ζ − λ) dζ
2p(xp y p ) 0
Z 1  1+ 1p   1p  q  1q
2 xp y p ′ xp y p
× F dζ
0 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1  1r
r
+ (2ζ − 2 + λ) dζ
1
2

1 1+ 1p   1p  q  1q 
xp y p xp y p
Z 
′
× F dζ
1
2
ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p
Z 1  1r
(y p − xp ) 2 r
= (2ζ − λ) dζ
2p(xp y p ) 0
Z 1  q+ pq   1p  q  1q
2 xp y p ′ xp y p
× F dζ
0 ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p

34
 1  1r Z 1  q+ pq  1p  q  1q 
xp y p xp y p
Z 
r ′
+ (2ζ − 2 + λ) dζ × F dζ .
1
2
1
2
ζxp + (1 − ζ)y p ζxp + (1 − ζ)y p

Since |F′ (a)|q isiin SR(ph), weihave

Z 1  1r Z 1  q+ pq
(y p − xp ) 2 r 2 xp y p
≤ (2ζ − λ) dζ
2p(xp y p ) 0 0 ζxp + (1 − ζ)y p
 l  1q
′ q ′ q 1 1
× h(1 − ζ) F (x) + h(ζ) F (y) − χϕ1 (ζ) p − p dζ
y x
Z 1  1r Z 1  q+ pq
r xp y p
+ (2ζ − 2 + λ) dζ
1
2
1
2
ζxp + (1 − ζ)y p
 l  1q 
′ q q 1 1
× h(1 − ζ) F (x) + h(ζ) (y) − χϕ1 (ζ) p − p dζ
y x
Z 1  1r
(y p − xp ) 2 r
≤ (2ζ − λ) dζ
2p(xp y p ) 0
q q
1q
× C9 (q, p; x, y) F′ (x) + C11 (q, p; x, y) F′ (y) + C13 (q, p; x, y)χ
Z 1  1r
2 r
+ (2ζ − 2 + λ) dζ
0

′ q ′

1
q
× C12 (q, p; y, x) F (x) + C10 (q, p; y, x) F (y) + C14 (q, p; y, x)χ q
1
(y p − xp ) λr+1 + (1 − λ)r+1 r 

q q
≤ × C9 (q, p; x, y) F′ (x) + C11 (q, p; x, y) F′ (y)
2p(xp y p ) 2(r + 1)

1q q q
1 
+ C13 (q, p; x, y)χ + C12 (q, p; y, x) F′ (x) + C10 (q, p; y, x) F′ (y) + C14 (q, p; y, x)χ q ,

hence we get the required result.

Remark 6. In Theorem 2.5.4 inserting h(ζ) = ζ, χ = 0, and l = 2 with ϕ1 (ζ) = ζ(1−ζ),

weiobtain Theoremi2.5 in reference [48].

Foriλ = 0, Theoremi2.5.4 reducesito theifollowingiresult.

Corollary 2.5.5. [44] LetiI = [α1 , β1 ] ⊂ R\{0} beiaip-harmonic convexiset, andilet

35
F : I = [α1 , β1 ] ⊆ R\{0} → Ribe aidifferentiableifunction onithe interior I o ofiI.
IfiF′ ∈ L[x, y] is in SR(ph) oniI, ir, q > 1, 1
r
+ 1
q
= 1,iand λ ∈ [0, 1], theni

1 
2xp y p p p(xp y p ) y F(a)
 Z
F − p da
xp + y p y − xp x a1+p
 1r
(y p − xp )

1 q
C9 (q, p; x, y) F′ (x)

≤ p p
×
2p(x y ) 2(r + 1)
q
1 (2.45)
+ C11 (q, p; x, y) F′ (y) + C14 (q, p; x, y)χ q
q q
+ C12 (q, p; y, x) F′ (x) + C10 (q, p; y, x) F′ (y)

1 
+ C14 (q, p; y, x)χ q ,

where C9 , C10 , C11 , C12 , C14 and C14 are given by (2.39)-(2.44).

Remark 7. In Corollary 2.5.5 inserting h(ζ) = ζ, χ = 0, and l = 2 with ϕ1 (ζ) =

ζ(1 − ζ), we obtain Corollary 3.6 in reference [48].

Theoremi2.5.4 reducesito theifollowing resultiby putting λ = 1.

Corollary 2.5.6. [44] Let I = [α1 , β1 ] ⊂ R\{0}ibe aip-harmoniciconvexiset, and let

F : I = [α1 , β1 ] ⊂ R\{0} → R beia differentiableifunction onithe interior I o iof I.iIf
F′ ∈ L[x, y],i|F′ |q iis a Strongly reciprocally (p, h)-convexifunction of higher orderion I,
1 1
r, q > 1, r
+ q
= 1, and λ ∈ [0, 1],itheni

F(x) + F(y) p(xp y p ) y F(a)

Z
− p da
2 y − xp x a1+p
 1r
(y p − xp )

1 q
C9 (q, p; x, y) F′ (x)

≤ p p
×
2p(x y ) 2(r + 1)
q
1
+ C11 (q, p; x, y) F′ (y) + C14 (q, p; x, y)χ q
q q
1 
+ C12 (q, p; y, x) F′ (x) + C10 (q, p; y, x) F′ (y) + C14 (q, p; y, x)χ q ,

36
where C9 , C10 , C11 , C12 , C13 and C14 are given by (2.39)-(2.44).

Remark 8. In Corollary 2.5.6 inserting h(ζ) = ζ, χ = 0, and l = 2 with ϕ1 (ζ) =

ζ(1 − ζ), we obtain Corollary 3.6 in reference [48].

For λ = 31 , Theorem 2.5.4 reducesito theifollowingiresult.

Corollary 2.5.7. [44] LetiI = [α1 , β1 ] ⊂ R\{0}ibe aip-harmoniciconvex set,iand

letiF : I = [α1 , β1 ] ⊂ R\{0} → Ribe aidifferentiable functionion theiinterior
I o iofiI.iIfiF′ ∈ L[x, y], i|F′ |q iisia Strongly reciprocallyi(p, h)-convexifunction of higher
orderioniI, r, q > 1,i 1r + 1
q
= 1, iandiλ ∈ [0, 1],ithen

1 
2xp y p p p(xp y p ) y F(a)
   Z
1
F(x) + 4F + F(y) − p da
6 xp + y p y − xp x a1+p
 1r
(y p − xp ) 1 + 2r+1

q q
C9 (q, p; x, y) F′ (x) + C11 (q, p; x, y) F′ (y)

≤ p p
× r
2p(x y ) 6.3 (r + 1)

1 q q
1 
+ C14 (q, p; x, y)χ q + C12 (q, p; y, x) F′ (x) + C10 (q, p; y, x) F′ (y) + C14 (q, p; y, x)χ q ,

where C9 , C10 , C11 , C12 , C13 and C14 are given by (2.39)-(2.44).

Remark 9. In Corollary 2.5.7 inserting h(ζ) = ζ, χ = 0, and l = 2 with ϕ1 (ζ) =

ζ(1 − ζ), we obtain Corollary 3.7 in reference [48].

For λ = 21 , Theorem 2.5.4 reducesito theifollowingiresult.

Corollary 2.5.8. [44] LetiI = [α1 , β1 ] ⊂ R\{0}ibe aip-harmoniciconvexiset, andilet

F : I = [α1 , β1 ] ⊂ R\{0} → R beia differentiableifunction onithe interioriI o iofiI.
IfiF′ ∈ L[x, y], |F′ |q isia Stronglyireciprocally (p, h)-convexifunction of higher order
oniI, r, q > 1,i 1r + 1
q
= 1,iand λ ∈ [0, 1],ithen

37
 1 
2xp y p p p(xp y p ) y F(a)
   Z
1
F(x) + 2F + F(y) − p da
4 xp + y p y − xp x a1+p
 1r
(y p − xp )

2 q
C9 (q, p; x, y) F′ (x)

≤ p p
× r
2p(x y ) 4.2 (r + 1)
q
1
+ C11 (q, p; x, y) F′ (y) + C14 (q, p; x, y)χ q
q q
1 
+ C12 (q, p; y, x) F′ (x) + C10 (q, p; y, x) F′ (y) + C14 (q, p; y, x)χ q ,

where C9 , C10 , C11 , C12 , C13 and C14 are given by (2.39)-(2.44).

Remark 10. In Corollary 2.5.8, inserting h(ζ) = ζ, χ = 0, and l = 2 with ϕ1 (ζ) =

ζ(1 − ζ), we obtain Corollary 3.8 in reference [48].

38
Chapter 3

Main Results

This chapter began with defining h-convexifunction, Godunova–Levin function, s-

convexifunction in the second sense, and p-convexifunction. In Section 3.2 we discussed
a newly defined mapping Mw
F (x) for h-convexifunction, see [32], a couple of basic re-

sults in the form of lemma and proposition are made. In Section 3.2.1 we utilize these
results to establishia newigeneralized Fejér-typeiinequality. Afterward, withisome mod-
ifications, we construct a generalizediform ofiHermite-Hadamardiinequality in Section
3.2.2. In the end, we talked about the most recent mapping that has been defined as

F (x) for h-convexifunction.

Hw

3.1 Introduction

As, h-convexifunction is also known as SX(h, I), the classes such asiGodunova–Levinifunctions
knownias Q(I) [49], s-convexifunctions inithe secondisense knownias Ks2 [50], and
p-convexifunctions knownias P (I) [51]. Here we have some basic definitions of the
Godunova–Levin function and s-convexifunction.

39
Definition 3.1.1 (Godunova–Levin function). Consider a non-negative con-
vexifunction F : I ⊆ R → R and foriall a, b ∈ Iiand ζ ∈ (0, 1), weihave theifollowing
inequality:

F(a) F(b)
F ζa + (1 − ζ)b ≤ + .
ζ 1−ζ

Definition 3.1.2 (s-convexifunction). Consider ainon-negative convexifunctioniF :

I ⊆ R → Riand foriall a, b ∈ I,is ∈ (0, 1], and ζ ∈ [0, 1],iwe haveithe follow-
ingiinequality:
F ζa + (1 − ζ)b ≤ ζ s F(a) + (1 − ζ)s F(b).

From the Definition (2.1.11), we consider h(ζ) = ζ, theniall non-negativeiconvex

functionsibelong toiSX(h, I). Ifithe aboveiinequality isireversed F isisaid toibe
h-concaveior F ∈SV(h, I). Moreover, allinon-negativeiconcave functionsibelongs
toiSV(h, I) for h(ζ) = ζ. We can included Q(I), Ks2 , and P (I) in theiclass of
h-convexifunctions.iIf h(ζ) = 1
ζ
, h(ζ) = 1, andih(ζ) = ζ s , where s ∈ (0, 1),ithen
Q(I) = SX(h, I),iP (I) ⊆ SX(h, I),iand Ks2 ⊆ SX(h, I),irespectively.
Onward inithis thesis,ithe function hiis considered integrableion [0, 1].

3.2 The mapping Mw

F (x)

Consider twoireal numbersiκ1 < κ2 , consideriintegrable functionsiF : [κ1 , κ2 ] = I →

Riand w : [κ1 , κ2 ] = I → R+ ∪ {0}. Defineithe mappingiMw
F (x) : [0, 1] → Rias

Z Cx (µ,ν) Z κ2
Mw
F (x) = F(a)w(a) da + F(a)w(a) da,
κ1 Dx (µ,ν)

where
 
Cx (µ, ν) = min µ(x), ν(x) , Dx (µ, ν) = max µ(x), ν(x) ,

40
forix ∈ [0, 1] weidefine µ : [0, 1] → [κ1 , κ2 ] and ν(x) : [0, 1] → [κ1 , κ2 ] as

µ(x) = xκ2 + (1 − x)κ1 , ν(x) = xκ1 + (1 − x)κ2

Note that, foriw = 1iin Mw

F (x) we have

Z Cx (µ,ν) Z κ2
M1F (x) = F(a) da + F(a) da,
κ1 Dx (µ,ν)

Theifollowing lemmaiwill beiuseful toius often.

Lemma 3.2.1. [32] Consideritwo functionsiF : [κ1 , κ2 ] → Riand w : [κ1 , κ2 ] → R+ ∪

{0}.iAlso,iforiany s ∈ [0, 1], idefineithe iϕs : [κ1 , κ2 ] × [κ1 , κ2 ] → [κ1 , κ2 ]iasiϕs (a, b) =
sa + (1 − s)biforia, b ∈ [κ1 , κ2 ].iThenifor allix ∈ [0, 1],

(i)
Cx (µ, ν) + Dx (µ, ν) = µ(x) + ν(x) = κ1 + κ2 .

(ii)
Dx (µ, ν) − Cx (µ, ν) = |µ(x) − ν(x)| = |1 − 2x|(κ2 − κ1 ).

(iii)
F(Cx (µ, ν)) + F(Dx (µ, ν)) = [F ◦ µ](x) + [F ◦ ν](x).

(iv)
ϕs (Cx (µ, ν), Dx (µ, ν)) + ϕs (Dx (µ, ν), Cx (µ, ν)) = κ1 + κ2 .

κ1 +κ2
Ifiwiis symmetricion [κ1 , κ2 ]iwith respectito 2
, ithen:

(v) Itiis symmetricion theiinterval [Cx (µ, ν), Dx (µ, ν)]iwithirespect toi κ1 +κ
2
2
,

41
 (vi) Weihave theifollowingiidentities:

[w ◦ ϕs ] Cx (µ, ν), Dx (µ, ν) = [w ◦ ϕs ] Dx (µ, ν), Cx (µ, ν)



= [w ◦ ϕs ] µ(x), ν(x) = [w ◦ ϕs ] ν(x), µ(x) ,

(vii) Weihave theifollowing integraliinequalities:

κ1 +κ2
Z Cx (µ,ν) Z κ2 Z
2
Z Dx (µ,ν)
w(a) da = w(a) da, w(a) da = w(a) da.
κ1 +κ2
κ1 Dx (µ,ν) Cx (µ,ν) 2

Proof. Theiproofsiof (i), (iii),iandi(iv) areistraight-forward,iand fori(ii), weican uti-

lizeithe followingiidentities

a + b − |b − a|
min{a, b} = ,
2

and
a + b + |b − a|
max{a, b} = .
2

Fori(v),isuppose κ1 ≤ Cx (µ, ν) ≤ κ1 +κ2

2
≤ Dx (µ, ν) ≤ κ2 ,iand soifor a ∈
[Cx (µ, ν), Dx (µ, ν)],ifrom (i)iweihave

w Cx (µ, ν) + Dx (µ, ν) − a = w(κ1 + κ2 − a) = w(a),

since,iwiis symmetriciwith respectito κ1 +κ2

2
.iFori(vi), we musticonsider theifollowingiequalities:

[w ◦ ϕs ] ν(x), µ(x) = w s xκ2 + (1 − x)κ1 + (1 − s) xκ1 + (1 − x)κ2




= w x sκ2 + (1 − s)κ1 + (1 − x) sκ1 + (1 − s)κ2

42




= w (1 − x) sκ2 + (1 − s)κ1 + x sκ1 + (1 − s)κ2




= w s xκ1 + (1 − x)κ2 + (1 − s) xκ2 + (1 − x)κ1


= [w ◦ ϕs ] µ(x), ν(x) .

Finally,iit isisufficient toiuse (i) and theichange ofivariableiu = κ1 + κ2 − a fori(vii).

ByiLemma 3.2.1iwe haveisome basiciproperties forithe mappingiMw

F (x)ibelow.

Proposition 6. [32] Consideritwo functionsiF : [κ1 , κ2 ] → Riandiw : [κ1 , κ2 ] → R+ ∪

{0}.iThen:i

(i)
Mw w
F (x) = MF (1 − x) ∀x ∈ [0, 1],

whichishowsithat Mw
F (x)iis symmetricion [κ1 , κ2 ] withirespect toi 2 .
1

(ii) Forisymmetriciwion [κ1 , κ2 ]iwithirespectito κ1 +κ2

2
iandip, q ≥ 1iwithi p1 + 1
q
=
1,iweihave
Mw
F (x) ≤ ∥F∥p ∥w∥q .

Also,iif Cx (µ, ν) = µ(x),ithen

  1q
1  1
Mw
F (x) ≤ x(κ2 − κ1 ) p ∥w∥q ∥F∥∞ ,
2

andiif Cx (µ, ν) = ν(x),ithen

  1q
1  1
Mw
F (x) ≤ (1 − x)(κ2 − κ1 ) p ∥w∥q ∥F∥∞ .
2

(iii) Let,ithe functioni(Fw)(a) = F(a)w(a)iis convexion [κ1 , κ2 ].iIf Cx (µ, ν) =

43
 MwF (x)
ν(x)iforisome x ∈ [0, 1),ithen theifunctioni 1−x
iisiconvex.iAlso, if Cx (µ, ν) =
Mw
F (x)
µ(x) forisome x ∈ (0, 1],ithen theifunctioni x
isiconvex.

(iv) Considering F and w are two continuous functions on [κ1 , κ2 ]. If F is

F (x)iis increas-
non-negative (non-positive) oni[κ1 , κ2 ], thenithe functioniMw
ingi(decreasing) oni[0, 21 )iand isidecreasing (increasing)ion ( 12 , 0].iAlso, Mw
F (x)

has a relative extremeipoint atix = 1

2
.iIf w ̸≡ 0,ithen correspondingito
anyia ∈ [κ1 , κ2 ] \ { κ1 +κ
2
2
}isatisfying

F(a) + F(κ1 + κ2 − a) = 0,

thereiexists aicritical pointifor Mw

F (x).i

Proof. (i) Thisifollows from theifacts µ(1 − x) = ν(x)iandiν(1 − x) = µ(x).

(ii) SinceiCx (µ, ν) ≤ κ1 +κ2

2
≤ Dx (µ, ν),iusing the statementsi(iii) andi(vii) iniLemma
3.2.1 and Hölder’s inequality, weihave theifollowing inequalities:

Mw
F (x)
Z Cx (µ,ν)  p1 Z Cx (µ,ν)  1q
p q
≤ F(a) da w(a) da
κ1 κ1
Z κ2  p1 Z κ2  1q
p q
+ F(a) da w(a) da
Dx (µ,ν) Dx (µ,ν)
Z Cx (µ,ν)  1q Z Cx (µ,ν)  p1 Z κ2  p1 
q p p
= w(a) da F(a) da + F(a) da
κ1 κ1 Dx (µ,ν)
Z Cx (µ,ν)  1q Z Cx (µ,ν)  p1 Z κ2  p1 
1 q p p
≤ w(a) da F(a) da + F(a) da .
2 κ1 κ1 Dx (µ,ν)

Thusi(ii) isiproved.

44
(iii) Weiprove theifirst part.iConsidering theichanges ofivariable a = xκ1 + (1 −
MwF (x)
x)uiandia = xκ2 + (1 − x)uiin twoiintegrals ofiH(x) = 1−x
, we obtain that

Z κ2

Z κ2

H(x) = (Fw) xκ1 + (1 − x)a da + (Fw) xκ2 + (1 − x)a da.
κ1 κ1

Now for x1 , x2 ∈ [0, 1) and non-negative r, s with r + s = 1, we have

Z κ2


H(rx1 + sx2 ) = (Fw) (rx1 + sx2 )κ1 + 1 − (rx1 + sx2 ) a da
κ1
Z κ2


+ (Fw) (rx1 + sx2 )κ2 + 1 − (rx1 + sx2 ) a da
κ1
κ2
Z



Z κ2



= (Fw) r x1 κ1 + (1 − x1 )a da + (Fw) s x2 κ1 + (1 − x2 )a da
κ1 κ1
Z κ2 Z κ2





+ (Fw) r x1 κ2 + (1 − x1 )a da + (Fw) s x2 κ2 + (1 − x2 )a da
Zκ1κ2 κ
Z κ2 1 



≤r (Fw) x1 κ1 + (1 − x1 )a da + (Fw) x1 κ2 + (1 − x1 )a da
κ1 κ1
Z κ2 Z κ2 



+s (Fw) x2 κ1 + (1 − x2 )a da + (Fw) x2 κ2 + (1 − x2 )a da
κ1 κ1

= rH(x1 ) + sH(x2 ).

(iv) Using theifact thatiwiis symmetricion [κ1 , κ2 ] withirespect toi κ1 +κ

2
2
and the Leib-
niz integral rule, weiobtain theifollowingiresult:

x ∈ [0, 12 ),

1 dMw [w ◦ µ](x){[F ◦ µ](x) + [F ◦ ν](x)},

F
(x) =
κ2 − κ1 dx
−[w ◦ ν](x){[F ◦ µ](x) + [F ◦ ν](x)}, x ∈ ( 1 , 1).


2

Proposition 7. LetiF : [κ1 , κ2 ] → R, iG : [κ1 , κ2 ] → R, andiw : [κ1 , κ2 ] →

R+ ∪ {0}ibeiintegrable functions. The mappings Mw
F (x) and Mg (x)ifor twoireal
w

45
numbersiκ1 ≤ κ2 .iDefine theimapping Mw
F (x) : [0, 1] → Rias:

Z Cx (µ,ν) Z κ2
Mw
F (x) = F(a)w(a) da + F(a)w(a) da,
κ1 Dx (µ,ν)

and Mw
g (x) : [0, 1] → Rias

Z Cx (µ,ν) Z κ2
Mw
g (x) = g(a)w(a) da + g(a)w(a) da.i
κ1 Dx (µ,ν)

TheniF + g : [κ1 , κ2 ] → Riis alsoiintegrable function.

Proof.

Z Cx (µ,ν) Z κ2
Mw
F+g (x) = (F + g)(a)w(a) da + (F + g)(a)w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2
= [(F(a) + g(a)]w(a) da + [(F(a) + g(a)]w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2
= [F(a)w(a) + g(a)w(a)] da + [F(a)w(a) + g(a)w(a)] da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z Cx (µ,ν)
= F(a)w(a) da + g(a)w(a) da
κ1 κ1
Z κ2 Z κ2
+ F(a)w(a) da + g(a)w(a) da
Dx (µ,ν) Dx (µ,ν)
Z Cx (µ,ν) Z κ2
= F(a)w(a) da + F(a)w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2
+ g(a)w(a) da + g(a)w(a) da
κ1 Dx (µ,ν)

= Mw w
F (x) + Mg (x).

Hence proved.

Proposition 8. LetiF : [κ1 , κ2 ] → R,iand w : [κ1 , κ2 ] → R+ ∪ {0} beitwo inte-

grableifunctions. Thenifor anyiλ ≥ 0,iλF : [κ1 , κ2 ] → R, iisialso integrable function.

46
Defineithe mappingiMw
F (x) : [0, 1] → Rias:

Z Cx (µ,ν) Z κ2
Mw
F (x) = F(a)w(a) da + F(a)w(a) da.
κ1 Dx (µ,ν)

Proof. Let λ ≥ 0,
Considering the mapping Mw
λF (x)

Z Cx (µ,ν) Z κ2
Mw
λF (x) = (λF(a))w(a) da + (λF(a))w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2
= λF(a)w(a) da + λF(a)w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2
=λ F(a)w(a) da + λ F(a)w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2 
=λ F(a)w(a) da + F(a)w(a) da = λMw
F (x).
κ1 Dx (µ,ν)

Hence proved.

Proposition 9. Theimapping Mw
F (x)iforitwo realinumbersiκ1 ≤ κ2 integrableifunctions

Fi : [κ1 , κ2 ] → R, and w : [κ1 , κ2 ] → R+ ∪ {0}, i1 ≤ i ≤ n.iForiλi ≥ 0, i1 ≤ i ≤ nithe

function F : [0, 1] → R where F = ni=1 λi Fi is also integrable function.
P

Proof. Considering the mappingiMw

F (x)

Z Cx (µ,ν) Z κ2
Mw
F (x) = F(a)w(a) da + F(a)w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) n
X Z κ2 n
X
= λi Fi (a)w(a) da + λi Fi (a)w(a) da
κ1 i=1 Dx (µ,ν) i=1
Z Cx (µ,ν)  
= λ1 F1 (a) + λ2 F2 (a) + · · · + λn Fn (a) w(a) da
κ1
Z κ2  
+ λ1 F1 (a) + λ2 F2 (a) + · · · + λn Fn (a) w(a) da
Dx (µ,ν)

47
 Z Cx (µ,ν)  
= λ1 F1 (a)w(a) + λ2 F2 (a)w(a) + · · · + λn Fn (a)w(a) da
κ1
Z κ2  
+ λ1 F1 (a)w(a) + λ2 F2 (a)w(a) + · · · + λn Fn (a)w(a) da
Dx (µ,ν)
Z Cx (µ,ν) Z Cx (µ,ν) Z Cx (µ,ν)
= λ1 F1 (a)w(a) da + λ2 F2 (a)w(a) da + · · · + λn Fn (a)w(a) da
κ1 κ1
Z κ2 Z κ2 Zκ1κ2
+ λ1 F1 (a)w(a) da + λ2 F2 (a)w(a) da + ··· + λn Fn (a)w(a) da
Dx (µ,ν) Dx (µ,ν) Dx (µ,ν)
Z Cx (µ,ν) Z Cx (µ,ν) Z Cx (µ,ν)
= λ1 F1 (a)w(a) da + λ2 F2 (a)w(a) dx + · · · + λn Fn (a)w(a) da
κ1 κ1
Z κ2 Z κ2 Zκ1κ2
+ λ1 F1 (a)w(a) da + λ2 F2 (a)w(a) dx + · · · + λn Fn (a)w(a) da
Dx (µ,ν) Dx (µ,ν) Dx (µ,ν)
Z Cx (µ,ν) Z κ2 
= λ1 F1 (a)w(a) da + F1 (a)w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2 
+ λ2 F2 (a)w(a) da + F2 (a)w(a) dx
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2 
+ · · · + λn Fn (a)w(a) da + Fn (a)w(a) da
κ1 Dx (µ,ν)

= λ1 Mw w w
F1 (x) + λ2 MF2 (x) + · · · + λn MFn (x)
n
X
= λi M w
Fi (x).
i=1

Hence proved.

Proposition 10. Let Fi : [κ1 , κ2 ] → R, i1 ≤ i ≤ niandiw : [κ1 , κ2 ] → R+ ∪ {0} be

integrable functions. Then F = max Fi , i = 1, 2, . . . , n is also integrable function.

Proof.

Z Cx (µ,ν) Z κ2
Mw
F (x) = max Fi (a)w(a) da + max Fi (a)w(a) da
κ1 Dx (µ,ν)
Z Cx (µ,ν) Z κ2
= Fc (a)w(a) da + Fc (a)w(a) da = Mw
Fc (x).
κ1 Dx (µ,ν)

Hence proved.
48
3.2.1 A New Generalized form of Fejér’s inequality

Here is a new and improved generalization of Fejér’s inequality, extending its applica-

F (x) for h-convex functions. This will broaden

bility to the newly defined mapping Mw
the scope and provide a more flexible tool for analysis. This is also known as Hermite-
Hadamard-Fejér’s inequality for an h-convexifunction. Once the derivation is done, we
developed several resultsiin theiform oficorollaries and remarks, for deeper insights into
the inequality’s structure.

Theorem 3.2.2. Considering h-convexifunctioniF : [κ1 , κ2 ] = I → Riandiw :

[κ1 , κ2 ] → R+ ∪ {0}isuch thatiw isisymmetric withirespect toi κ1 +κ
2
2
.iFor allix ∈
[0, 1],iwe haveithe followingiinequality:

  Z Dx (µ,ν)
1 κ1 + κ2
F w(a) da
2h( 12 ) 2 Cx (µ,ν)
Z κ2
≤ F(a)w(a) da − Mw F (x)
κ1
 µ(x)  (3.1)
|ν(x) − µ(x)|[F ◦ µ](x) + [F ◦ ν](x) a − ν(x)
Z
≤ h w(a) da
(µ(x) − ν(x)) ν(x) µ(x) − ν(x)
|ν(x) − µ(x)|[F ◦ µ](x) + [F ◦ ν](x) ν(x)
 
a − µ(x)
Z
= h w(a) da.
(ν(x) − µ(x)) µ(x) ν(x) − µ(x)

Proof. For any ρ ∈ [Cx (µ, ν), Dx (µ, ν)] existsik ∈ (0, 1) suchithatiρ = kCx (µ, ν) +
k1 Dx (µ, ν), k1 = 1 − k. From theidefinition ofian h-convexifunction, we have

F(kCx (µ, ν) + k1 Dx (µ, ν))w(kCx (µ, ν) + k1 Dx (µ, ν))

≤ h(k)F Cx (µ, ν) + h(k1 )F(Dx (µ, ν)) w(kCx (µ, ν) + k1 Dx (µ, ν))

= h(k)F Cx (µ, ν)w(kCx (µ, ν) + k1 Dx (µ, ν)) + h(k1 )F(Dx (µ, ν))w(kCx (µ, ν) + k1 Dx (µ, ν)),
(3.2)

49
and

F(k1 Cx (µ, ν) + kDx (µ, ν))w k1 Cx (µ, ν) + kDx (µ, ν)



≤ h(k1 )F Cx (µ, ν) + h(k)F Dx (µ, ν)) w(k1 Cx (µ, ν) + kDx (µ, ν))


= h(k1 )F Cx (µ, ν) w(k1 Cx (µ, ν) + kDx (µ, ν)) + h(k)F Dx (µ, ν))w(k1 Cx (µ, ν) + kDx (µ, ν)).
(3.3)
After adding (3.2) and (3.3), and integrating with respect to k over [0, 1], we obtain

Z 1


F kCx (µ, ν) + k1 Dx (µ, ν) w kCx (µ, ν) + k1 Dx (µ, ν) dk
0
Z 1



+ F k1 Cx (µ, ν) + kDx (µ, ν) w k1 Cx (µ, ν) + kDx (µ, ν) dk
0
Z 1



≤ h(k)F Cx (µ, ν) w kCx (µ, ν) + k1 Dx (µ, ν)
0




+ h(k1 )F Dx (µ, ν) w kCx (µ, ν) + k1 Dx (µ, ν) dk
Z 1



+ h(k1 )F Cx (µ, ν) w k1 Cx (µ, ν) + kDx (µ, ν)
0




+ h(k)F Dx (µ, ν) w k1 Cx (µ, ν) + kDx (µ, ν) dk.

Considering leftihand sideiof theiinequality and substituting a = kCx (µ, ν) +

k1 Dx (µ, ν), a = k1 Cx (µ, ν) + kDx (µ, ν) in the first and second integral respectively, we
get

Z Cx (µ,ν) Z Dx (µ,ν)
1 1
F(a)w(a) da + F(a)w(a) da
Cx (µ, ν) − Dx (µ, ν) Dx (µ,ν) Dx (µ, ν) − Cx (µ, ν) Cx (µ,ν)
Z Dx (µ,ν)
2
= F(a)w(a) da,
Dx (µ, ν) − Cx (µ, ν) Cx (µ,ν)
(3.4)

50
solving theiright-handiside ofithe inequalityi

Z 1



F Cx (µ, ν) h(k)w kCx (µ, ν) + k1 Dx (µ, ν) + h(k1 )w k1 Cx (µ, ν) + kDx (µ, ν) dk
0
Z 1




+ F Dx (µ, ν) h(k1 )w kCx (µ, ν) + k1 Dx (µ, ν) + h(k)w k1 Cx (µ, ν) + Dx (µ, ν) dk
0

1
Z


= 2F Cx (µ, ν) h(s)w sCx (µ, ν) + (1 − s)Dx (µ, ν) ds
0

1
Z


+ 2F Dx (µ, ν) h(s)w (1 − s)Cx (µ, ν) + sDx (µ, ν) ds
0

 1
Z


= 2 F Cx (µ, ν) + F Dx (µ, ν) h(s)w(sCx (µ, ν) + (1 − s)Dx (µ, ν)) ds
0

 1
Z

 

= 2 F Cx (µ, ν) + F Dx (µ, ν) h(s) w ◦ ϕs Cx (µ, ν), Dx (µ, ν) ds,
0
(3.5)
here we use the symmetry of the weight w. Now combining and simplifying the results
obtained in equation (3.4) and (3.5)

Z Dx (µ,ν)
1
F(a)w(a) da
Dx (µ, ν) − Cx (µ, ν) Cx (µ,ν)
(3.6)

 1
Z


≤ F Cx (µ, ν) + F Dx (µ, ν) h(s)[w ◦ ϕs ](Cx (µ, ν), Dx (µ, ν)) ds,
0

Let F be an h-convexifunction

F(ju + (1 − j)v)) ≤ h(j)F(u) + h(1 − j)F(v).

Now substituting j = 21 , u = kCx (µ, ν) + (1 − k)Dx (µ, ν) and v = (1 − k)Cx (µ, ν) +

kDx (µ, ν), simplifying the equation

 
Cx (µ, ν) + Dx (µ, ν)
F
2
1  
≤ h( ) F(kCx (µ, ν) + (1 − k)Dx (µ, ν) + F((1 − k)Cx (µ, ν) + kDx (µ, ν) .
2

51
Now, multiply w kCx (µ, ν) + (1 − k)Dx (µ, ν) = w (1 − k)Cx (µ, ν) + kDx (µ, ν) and

integrate with respect to k over [0, 1]

 Z 1
Cx (µ, ν) + Dx (µ, ν)

F w kCx (µ, ν) + (1 − k)Dx (µ, ν) dk
2 0
Z 1
1


≤ h( ) F kCx (µ, ν) + (1 − k)Dx (µ, ν) w kCx (µ, ν) + (1 − k)Dx (µ, ν) dk
2 0
Z 1 



+ F (1 − k)Cx (µ, ν) + kDx (µ, ν) w (1 − k)Cx (µ, ν) + kDx (µ, ν) dk ,
0

after some suitable substitutions, we obtain

  Z Dx (µ,ν)
Cx (µ, ν) + Dx (µ, ν) 1
F w(a) da
2 Dx (µ, ν) − Cx (µ, ν) Cx (µ,ν)
2h( 12 )
Z Dx (µ,ν)
≤ F(a)w(a) da,
Dx (µ, ν) − Cx (µ, ν) Cx (µ,ν)

or

  Z Dx (µ,ν) Z Dx (µ,ν)
1 Cx (µ, ν) + Dx (µ, ν)
F w(a) da ≤ F(a)w(a) da. (3.7)
2h( 12 ) 2 Cx (µ,ν) Cx (µ,ν)

Nowicombine equation (3.6) and (3.7), we obtain

  Z Dx (µ,ν)
1 Cx (µ, ν) + Dx (µ, ν)
F w(a) da
2h( 12 ) 2 Cx (µ,ν)
Z Dx (µ,ν)
≤ F(a)w(a) da
Cx (µ,ν) (3.8)
 


≤ Dx (µ, ν) − Cx (µ, ν) F Cx (µ, ν) + F Dx (µ, ν)
Z 1
× h(s)[w ◦ ϕs ](Cx (µ, ν), Dx (µ, ν)) ds.
0

52
By the definition of mapping Mw
F foriall x ∈ [0, 1] weiget theifollowing identity

Z κ2 Z Dx (µ,ν)
F(a)w(a) da − F(a)w(a) da = Mw
F (x). (3.9)
κ1 Cx (µ,ν)

Now using equation (3.9) and (ii), (iii), (v), and (vi) statements of Lemma 3.2.1 we get
theifollowing result

  Z Dx (µ,ν)
1 κ1 + κ2
F w(a) da
2h( 12 ) 2 Cx (µ,ν)
Z κ2
≤ F(a)w(a) da − Mw F (x)
κ1
Z 1
(3.10)


≤ |ν(x) − µ(x)| [F ◦ µ](x) + [F ◦ ν](x) h(s)[w ◦ ϕs ](µ(x), ν(x)) ds
0


Z 1
= |ν(x) − µ(x)| [F ◦ µ](x) + [F ◦ ν](x) h(s)[w ◦ ϕs ](ν(x), µ(x)) ds.
0

Applyingithe changeiof variableib = ϕs µ(x), ν(x) ior b = ϕs ν(x), µ(x) inithe last

twoiintegralsiin (3.10) andiconsiderithat

µ(x)   Z ν(x)  
a − ν(x) a − µ(x)
Z
1 1
h w(a) da = h w(a) da.
µ(x) − ν(x) ν(x) µ(x) − ν(x) ν(x) − µ(x) µ(x) ν(x) − µ(x)
(3.11)
Finally, by using equations (3.10) and (3.11) weiget theidesirediresult.

The aboveitheorem is valid in general. Furthermore, the following corollary holds

for more specific values.

Corollary 3.2.3. Considering integrable functions defined as h : [0, max 1, κ2 − κ1 ],

F : [κ1 , κ2 ] → Riand w : [κ1 , κ2 ] → R+ ∪ {0}, w ≥ 0,iwhere hiis multiplicative or super
multiplicative, F is non-negative h-convexifunction, andiw isisymmetric withirespect

53
 R Dx (µ,ν)
toi κ1 +κ
2
2
and Cx (µ,ν)
w(x) dx > 0. Then

  Z κ2
κ1 + κ2
F ≤B F(a)w(a) da, (3.12)
2 κ1

 Dx (µ,ν)−Cx (µ,ν) 
C (µ,ν)+Dx (µ,ν)
2h( 1 ) h(u)w(u+ x
R
0 2 ) du
whereiB = min R Dx (µ,ν) 2 , Cx (µ,ν)+D x (µ,ν)
2
iand
Cx (µ,ν)
w(a) da R 2
R Dx (µ,ν)
Cx (µ,ν) Cx (µ,ν)+Dx (µ,ν)
h(v−u)w(v)w(u) dvdu
2
R Cx (µ,ν)+D x (µ,ν) R
Dx (µ,ν)
Cx (µ,ν)
2
Cx (µ,ν)+Dx (µ,ν) h(v − u)w(v)w(u) dvdu ̸= 0, h(u) ̸= 0 for u ̸= 0.
2

Proof. Let h be super-multiplicative h(u) ̸= 0iforiu ̸= 0. Then h(u) > 0iforiu > 0
weihave u, v ∈ [κ1 , κ2 ] such thatiκ1 ≤ Cx (µ, ν) ≤ u < Cx (µ,ν)+Dx (µ,ν)
2
< v ≤ Dx (µ, ν) ≤
κ2 weihave

Cx (µ,ν)+Dx (µ,ν)   Cx (µ,ν)+Dx (µ,ν)

v− −u
 
κ1 + κ2 Cx (µ, ν) + Dx (µ, ν) 2 2
= = u+ v.
2 2 v−u v−u

Cx (µ,ν)+Dx (µ,ν) Cx (µ,ν)+Dx (µ,ν)

v− −u
Denote δ = v−u
2
> 0. Then γ = 1 − δ = 2
v−u
and Cx (µ,ν)+Dx (µ,ν)
2
=
δu + γv, and F( Cx (µ,ν)+D
2
x (µ,ν)
) = F(δu + γv) ≤ h(δ)F(u) + h(γ)F(v). Since hiis super
 
C (µ,ν)+Dx (µ,ν) C (µ,ν)+Dx (µ,ν)
v− x h(v− x )
multiplicative, we have h(δ) = h v−u
2
≤ 2
h(v−u)
similarly
Cx (µ,ν)+Dx (µ,ν)
h( −u)
h(γ) ≤ 2
h(v−u)
. So, when F > 0 we have

h(v − Cx (µ,ν)+D x (µ,ν)

h( Cx (µ,ν)+Dx (µ,ν)
− u)
 
Cx (µ, ν) + Dx (µ, ν) 2
) 2
F ≤ F(u) + F(v),
2 h(v − u) h(v − u)

multiplying by h(v − u) on both sides

 
Cx (µ, ν) + Dx (µ, ν)
h(v − u)F
2
    (3.13)
Cx (µ, ν) + Dx (µ, ν) Cx (µ, ν) + Dx (µ, ν)
≤h v− F(u) + h − u f (v).
2 2

54
This inequality continues to hold when h is multiplicative, regardless the sign of F.

Multiplying equation (3.13) with w(u) and integrating over theiinterval

[Cx (µ, ν), Cx (µ,ν)+D
2
x (µ,ν)
] withirespect toi du, and we then multiply with w(v) and inte-
grate over intervali[ Cx (µ,ν)+D
2
x (µ,ν)
, Dx (µ, ν)] withirespect toi dv we get

  Z Dx (µ,ν) Z Cx (µ,ν)+Dx (µ,ν) 

Cx (µ, ν) + Dx (µ, ν) 2
F h(v − u)w(u) du w(v) dv
2 Cx (µ,ν)+Dx (µ,ν)
2
C x (µ,ν)
Z Dx (µ,ν)   Z Cx (µ,ν)+Dx (µ,ν)
Cx (µ, ν) + Dx (µ, ν) 2
≤ h v− w(v) dv F(u)w(u) du
Cx (µ,ν)+Dx (µ,ν)
2
2 Cx (µ,ν)
Z Dx (µ,ν) Z Cx (µ,ν)+Dx (µ,ν)  
2 Cx (µ, ν) + Dx (µ, ν)
+ F(v)w(v) dv h − u w(u) du.
Cx (µ,ν)+Dx (µ,ν)
2
C x (µ,ν) 2

Onithe rightihand sideiwe applyithe substitutioniv − Cx (µ,ν)+Dx (µ,ν)

2
= t in the first
integral andi Cx (µ,ν)+D
2
x (µ,ν)
− u = t in the second integral of the sum, we get

  Z Dx (µ,ν) Z Cx (µ,ν)+Dx (µ,ν)

Cx (µ, ν) + Dx (µ, ν) 2
F h(v − u)w(u)w(v) dudv
2 Cx (µ,ν)+Dx (µ,ν)
2
Cx (µ,ν)
Z Dx (µ,ν)−Cx (µ,ν)   Z Cx (µ,ν)+Dx (µ,ν)
2 Cx (µ, ν) + Dx (µ, ν) 2
≤ h(t)w t + dy F(u)w(u) du
0 2 Cx (µ,ν)
Z Dx (µ,ν)−Cx (µ,ν)   Z Dx (µ,ν)
2 Cx (µ, ν) + Dx (µ, ν)
+ h(t)w − t dt F(v)w(v) dv
0 2 Cx (µ,ν)+Dx (µ,ν)
2
Z Dx (µ,ν)−Cx (µ,ν)   Z Dx (µ,ν)
2 Cx (µ, ν) + Dx (µ, ν)
= h(t)w t + dt. F(u)w(u) du,
0 2 Cx (µ,ν)

where, in the first we apply the symmetry property of the function w over the
interval [Cx (µ, ν), Dx (µ, ν)], i.e. w( Cx (µ,ν)+D
2
x (µ,ν)
− t) = w( Cx (µ,ν)+D
2
x (µ,ν)
+ t) for
t ∈ [0, Dx (µ,ν)−C
2
x (µ,ν)
]. Now applying Lemma 3.2.1 andisubstituting u = a on theiright
sideiof theiinequalityirespectively, we get the desired result

55
   R Dx (µ,ν)−Cx (µ,ν) Cx (µ,ν)+Dx (µ,ν)

Z Dx (µ,ν)
κ1 + κ2 0
2
h(t)w t + 2
dt
F ≤R R Cx (µ,ν)+D x (µ,ν)
× F(a)w(a) da
2 Dx (µ,ν) C (µ,ν)
Cx (µ,ν)+Dx (µ,ν) Cx (µ,ν)
2
h(v − u)w(u)w(v) dudv x
2

or we can write

  R Dx (µ,ν)−Cx (µ,ν) Cx (µ,ν)+Dx (µ,ν)
Z Dx (µ,ν)

κ1 + κ2 0
2
h(u)w u + 2
du
F ≤R R Cx (µ,ν)+D x (µ,ν)
× F(a)w(a) da
2 Dx (µ,ν) C (µ,ν)
Cx (µ,ν)+Dx (µ,ν) Cx (µ,ν)
2
h(v − u)w(u)w(v) dudv x
2
Z Dx (µ,ν)
=B F(a)w(a) da.
Cx (µ,ν)

Remark 11. (i) Inequality (3.1) isireversed if Fiis h-concaveifunction in the theorem
(3.2.2).
R Dx (µ,ν) R Cx (µ,ν)+Dx (µ,ν)
(ii) If h is subimultiplicative, Cx (µ,ν)+Dx (µ,ν) Cx (µ,ν)
2
h(v−u)w(u)w(v) dudv ̸= 0,
2

h ≥ 0, and ifiFiisih-concave functionithen the inequality (3.12) is reversed, where

constant B changes from min → max.

Here is another way of presenting the above generalization of Fejér’s inequality.

Corollary 3.2.4. Supposeithat F : [κ1 , κ2 ] → Riis non-negative integrable h-

convexifunctioniand w : [κ1 , κ2 ] → R+ ∪ {0}iis also integrable function and
κ1 +κ2
symmetriciwith respectito 2
. Foriall x ∈ [0, 1], iweihave theifollowingiinequality:

  Z Dx (µ,ν)
1 κ1 + κ2
F w(a) da
2h( 12 ) 2 Cx (µ,ν)
Z κ2
≤ F(a)w(a) da − Mw F (x)
κ1
Dx (µ,ν)   (3.14)
[F ◦ µ](x) + [F ◦ ν](x) Dx (µ, ν) − a
Z
≤ H w(a) da
2 Cx (µ,ν) Dx (µ, ν) − Cx (µ, ν)
[F ◦ µ](x) + [F ◦ ν](x) Cx (µ,ν)
 
a − Cx (µ, ν)
Z
= H w(a) da.
2 Dx (µ,ν) Dx (µ, ν) − Cx (µ, ν)

56
Proof. Consider F and h-convexifunction and some statements in Lemma 3.2.1

  


κ1 + κ2 ϕs Cx (µ, ν), Dx (µ, ν) + ϕs Dx (µ, ν), Cx (µ, ν)
F =F
2 2
 
1 


≤h F ◦ ϕs Cx (µ, ν), Dx (µ, ν) + F ◦ ϕs Dx (µ, ν), Cx (µ, ν)
2
 
1 

≤h H(s) F(Cx (µ, ν) + F Dx (µ, ν)
2
 
1

=h H(s) [F ◦ µ](x) + [F ◦ ν](x) ,
2

where H(s) = h(s) + h(1 − s),is ∈ [0, 1]. Using the symmetric property of the h-
convexifunction, we have

Dx (µ,ν)   Z Dx (µ,ν)  
Dx (µ, ν) − a Dx (µ, ν) − a
Z
H w(a) da = 2 h w(a) da.
Cx (µ,ν) Dx (µ, ν) − Cx (µ, ν) Cx (µ,ν) Dx (µ, ν) − Cx (µ, ν)

Multiplyingibyi[w◦ϕs ] Cx (µ, ν), Dx (µ, ν) ifor all s, x ∈ [0, 1]iandithen integratingiwith

respectito s ∈ [0, 1]iwe get required generalization of Fejér’s inequality given in Equa-
tion (3.14).

Remark 12. Thus, it is clear that (3.1) and (3.14) areiequivalent,iandithe only change
isiin the manner of presentationiand theiresultingiconsequences. These generalizations
of Fejér’s inequality for h-convexifunction provided in these inequalities are valuable
contributions to the literature.

Remark 13. Substituting h(s) = sη (η ∈ (0, 1]), h(s) = 1

s
(s ∈ (0, 1)),ih(s) = 1,
andih(s) = s in (3.1) and (3.14), we obtain generalizations of Fejér’s inequality
foris-convexifunctions inithe secondisense,iGodunova-Levinifunctions, p-convex func-
tions,iand convexifunctions,irespectively.
Now, taking particular value h(s) = siin (3.1)iand (3.14), weihave theiinequalities as

57
below,irespectively

  Z Dx (µ,ν) Z κ2
κ1 + κ2
F w(a) da ≤ F(a)w(a) da − Mw F (x)
2 Cx (µ,ν) κ1

[F ◦ µ](x) + [F ◦ ν](x) ν(x)

Z


≤ a − µ(x) w(a) da
|ν(x) − µ(x)| µ(x)
Z µ(x)
[F ◦ µ](x) + [F ◦ ν](x)

= a − ν(x) w(a) da,
|µ(x) − ν(x)| ν(x)
(3.15)
and

  Z Dx (µ,ν) Z κ2
κ1 + κ2
F w(a) da ≤ F(a)w(a) da − Mw F (x)
2 Cx (µ,ν) κ1
(3.16)
[F ◦ µ](x) + [F ◦ ν](x) Dx (µ,ν)
Z
≤ w(a) da.
2 Cx (µ,ν)

Ainew formiof generalizediFejér-type inequalityiis introduced in inequality (3.15),

whereas inequality (3.16) is a straightforward generalization of the classical Fejér
inequality for convex functions.

Remark 14. Substituting x = 0, 1 in (Mw

F (0) = MF (1) = 0),ithen weiget theifollowing
w

Fejér-typeiinequality for h-convexifunction obtainediin equation (3.10):

  Z κ2 Z κ2
1 κ1 + κ2
F w(a) da ≤ F(a)w(a) da
2h( 21 ) 2 κ1 κ1
Z 1
≤ (κ2 − κ1 )[F(κ1 ) + F(κ2 )] h(s)w(sκ1 + (1 − s)κ2 ) ds,
0
(3.17)
similarly, with this assumption in equation (3.14), we get ainew h-convexiversion
ofiFejér inequality:

58
   Z κ2 Z κ2
1 κ1 + κ2
F w(a) da ≤ F(a)w(a) da
2h( 12 ) 2 κ1 κ1
F(κ1 ) + F(κ2 ) κ2
 
κ2 − a
Z
≤ H w(a) da (3.18)
2 κ1 κ2 − κ1
F(κ1 ) + F(κ2 ) κ2
 
a − κ1
Z
= H w(a) da.
2 κ1 κ2 − κ1

Now, we discuss different inequalities related to generalized Fejér-typeiinequality

for h-convexifunction. For different h(s), w, andiforiany n ∈ N ,in ≥ 3. Inequalities of
this nature can be found in [52, 53, 54] along with the references.

Remark 15. Now consider theifollowing forin ∈ N , n ≥ 3,

κ1 + (n − 1)κ1 κ2 + (n − 1)κ1 κ1 + (n − 1)κ2 n − 2

κ1 = ≤ ≤ + (κ1 − κ2 )
n n n n
κ1 + (n − 1)κ2
≤ ≤ κ2 ,
n

andi
 
κ1 + (n − 1)κ2 κ2 + (n − 1)κ1
w + − a = w(κ1 + κ2 − a) = w(a).
n n

This shows thatiw is symmetric on [ κ2 +(n−1)κ

n
1
+ κ1 +(n−1)κ2
n
] with respect to κ1 +κ2
2
.
so, in case x = 1
n
in equation (3.1), and (3.14), we obtain the following inequalities
respectively:
  Z κ1 +(n−1)κ2
1 κ1 + κ2 n

1 F w(a) da
2h( 2 ) 2 κ2 +(n−1)κ1
n
Z κ1 +(n−1)κ2
n
≤ F(a)w(a) da
κ2 +(n−1)κ1
n (3.19)
    
n−2 κ2 + (n − 1)κ1 κ1 + (n − 1)κ2
≤ (κ2 − κ1 ) F +F
n n n
Z 1  
κ2 + (n − 1)κ1 κ1 + (n − 1)κ2
× h(s)[w ◦ ϕs ] , ds,
0 n n

59
and

  Z κ1 +(n−1)κ2
1 κ1 + κ2 n

1 F w(a) da
2h( 2 ) 2 κ2 +(n−1)κ1
n
Z κ1 +(n−1)κ2
n
≤ F(a)w(a) da
κ2 +(n−1)κ1
n
κ2 +(n−1)κ1
κ1 +(n−1)κ2
κ1 +(n−1)κ2   
F +F κ2 − a
Z
n n
n n 1
≤ H − w(a) da.
2 κ2 +(n−1)κ1
n
n−2 κ2 − κ1 n
(3.20)
Inequalities obtained in (3.19) and (3.20) deal with manyiHermite-HadamardiandiFejér’-
type inequalities ofithis kind foriall n ≥ 3.i

Example 3.2.1. Letin = 3, 4 and h(s) = siin (3.20)ito getiFejér’-typeiinequalities:

  Z κ1 +2κ2 Z κ1 +2κ2
κ1 + κ2 3 3
F w(a) da ≤ F(a)w(a) da
2 κ2 +2κ1
3
κ2 +2κ1
 3 
κ2 +2κ1 κ1 +2κ2



F 3
+F 3
Z κ1 +2κ2
3
≤ w(a) da,
2 κ2 +2κ1
3

and

  Z κ1 +3κ2 Z κ1 +3κ2
κ1 + κ2 4 4
F w(a) da ≤ F(a)w(a) da
2 κ2 +3κ1
4
κ2 +3κ1
4
κ2 +3κ1 κ1 +3κ2 Z κ1 +3κ 2



F 4
+ F 4
4
≤ w(a) da.
2 κ2 +3κ1
4

3.2.2 AiNew Generalized formiof Hermite-Hadamardiinequality

Here is the new and improved generalization ofithe celebratediHermite-Hadamardiinequality

relateditoih-convexifunctions anditheir sub-classes asiwell (seeialso [55, 56, 57, 58]).

60
Now we consider the case w ≡ 1 in equation (3.1) and (3.14) to get our desired result:

  Z Dx (µ,ν)
1 κ1 + κ2 1
F ≤ F(a) da
2h( 21 ) 2 (κ2 − κ1 )|1 − 2x| Cx (µ,ν)
Z 1
≤ ([F ◦ µ](x) + [F ◦ ν](x)) h(s) da (3.21)
0
[F ◦ µ](x) + [F ◦ ν](x) 1
Z
= H(s) da.
2 0

Now with the same assumption in equation (3.17) and(3.18), we obtain theiHermite-
Hadamardiinequality related to h-convexifunctions

  Z κ2 Z 1
1 κ1 + κ2 1
F ≤ F(a) da ≤ [F(κ1 ) + F(κ2 )] h(s) ds. (3.22)
2h( 12 ) 2 κ2 − κ1 κ1 0

Remark 16. Thus, we conclude that the equations (3.21) and (3.22) areiequivalent,
andithe only change isiin the manner of presentationiand consequences.iThese gen-
eralizations of Hermite-Hadamardiinequality forih-convexifunction provide valuable
contributions to theiliterature.
Further, considering the equations (3.21) and (3.22),itaking h(s) = sη (η ∈
(0, 1]),ih(s) = 1
s
(s ∈ (0, 1)),ih(s) = 1,iand h(s) = s , we obtain generalized
Hermite-Hadamardiinequality foris-convexifunctions in theisecond sense,iGodunova-
Levinifunctions,ip-convexifunctions, andiconvexifunctions,irespectively.
For particular case, if weiconsider h(s) = s,iandix ∈ [0, 1]\{ 21 }, in equation (3.21) we
found the following inequality:

  Z Dx (µ,ν)
κ1 + κ2 1 1
F ≤ F(a) da ≤ ([F ◦ µ](x) + [F ◦ ν](x)),
2 (κ2 − κ1 )|1 − 2x| Cx (µ,ν) 2
(3.23)
which is dependent on variable x. That is another updated versioniofiHermite-

61
Hadamardiinequality.
Forianother particular case, whenix = 0, 1 we retrieve theiclassical form ofiHermite-
Hadamardiinequality.

Newigeneralizations ofiHermite-Hadamardiinequalities forih-convexifunction are

formed, for different h(s), and any n ∈ N ,in ≥ 3, weihave:

Remark 17. Considering a particular case w ≡ 1 and h(s) = s in equation (3.19), we

obtain
  Z κ1 +(n−1)κ2
κ1 + κ2 n n
F ≤ F(a) da
2 (κ2 − κ1 )(n − 2) κ2 +(n−1)κ 1
  n  
1 κ2 + (n − 1)κ1 κ1 + (n − 1)κ2
≤ F +F ,
2 n n
which is aiparticular case foriHermite-Hadamardiinequality dependingionin.

Remark 18. Forifixedix ∈ [0, 1]iandia ∈ [κ1 , κ2 ],ithe weight functioniw can be written
as
w(a) = (Dx (µ, ν) − a)(a − Cx (µ, ν)),

where, wiis symmetricion [Cx (µ, ν), Dx (µ, ν)]iwithirespect toi κ1 +κ

2
2
. By using Lemma
3.2.1, we obtain the following results

Dx (µ,ν)
|1 − 2x|3 (κ2 − κ1 )3
Z



Dx (µ, ν) − a a − Cx (µ, ν) ) da =
Cx (µ,ν) 6

and

Z 1 Z 1
h(s)[w ◦ ϕs ] µ(x), ν(x) ds = |1 − 2x|2 (κ2 − κ1 )2


h(s)s(1 − s) ds.
0 0

By using Equation (3.10), we obtained the followingiHermite-Hadamardiinequality

62
forih-convexifunctions:
 
1 κ1 + κ2
F
12h( 12 ) 2
Z Dx (µ,ν)
1


≤ F(a) Dx (µ, ν) − a a − Cx (µ, ν) da
|1 − 2x| (κ2 − κ1 )3
3
Cx (µ,ν)

1
Z
≤ [F ◦ µ](x) + [F ◦ ν(x)] h(s)s(1 − s) ds.
0

In the particular values for x = 0, 1 and h(s) = s, we get

  Z κ2
κ1 + κ2 6 F(κ1 ) + F(κ2 )
F ≤ 3
F(a)(κ2 − a)(a − κ1 ) da ≤ ,
2 (κ2 − κ1 ) κ1 2

which was obtained from equation (3.17) .

3.3 The mapping Hw

F (x):

Consider twoireal numbersiκ1 < κ2 ,iconsider integrableifunctionsiF : [κ1 , κ2 ] = I ⊆

R → Riandiw : [κ1 , κ2 ] = I → R+ ∪ {0}.iDefine theimappingiHw
F (x) : [0, 1] → Rias

Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)

Hw
F (x) = F xa + (1 − x) w(a) da
κ1 2
Z κ2  
Cx (µ, ν) + Dx (µ, ν)

+ F xa + (1 − x) w(a) da,
Dx (µ,ν) 2

where
 
Cx (µ, ν) = min µ(x), ν(x) , Dx (µ, ν) = max µ(x), ν(x) ,

for x ∈ [0, 1] we defineiµ(x) : [0, 1] → [κ1 , κ2 ]iand ν(x) : [0, 1] → [κ1 , κ2 ]ias:

µ(x) = xκ2 + (1 − x)κ1 , i ν(x) = xκ1 + (1 − x)κ2 .

63
Note that,ifor w = 1 iniHw
F (x) we have

Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)

H1F (x) = F xa + (1 − x) da
κ1 2
Z κ2 
Cx (µ, ν) + Dx (µ, ν)

+ F xa + (1 − x) da.
Dx (µ,ν) 2

F (x) yields similar results to those established for the

Remark 19. The Mapping Hw

F (x) in Lemmai3.2.1
mapping Mw

Weiobtain some basiciproperties forithe mappingiHw

F (x) in the followingiproposition.

Proposition 11. Consideritwo integrableifunctionsiF : [κ1 , κ2 ] → Riandiw :

[κ1 , κ2 ] → R+ ∪ {0}.iThen

(i)
Hw w
F (x) = HF (1 − x) ∀x ∈ [0, 1],

whichishows thatiHw
F (x)iisisymmetric oni[κ1 , κ2 ] withirespect toi 2 .
1

(ii) Forisymmetriciw oni[κ1 , κ2 ] withirespect toi κ1 +κ

2
2
iandip, q ≥ 1 withi p1 + 1q = 1,iwe
havei
Hw
F (x) ≤ ∥F∥p ∥w∥q .

Also,iifiCx (µ, ν) = µ(x), itheni

  1q
1  1
Hw
F (x) ≤ x(κ2 − κ1 ) p ∥w∥q ∥F∥∞ ,
2

andiif Cx (µ, ν) = ν(x),ithen

  1q
1  1
Hw
F (x) ≤ (1 − x)(κ2 − κ1 ) p ∥w∥q ∥F∥∞ .
2

64
(iii) Let, theifunction (Fw)(a) = F(a)w(a)iis convexioni[κ1 , κ2 ].iIf Cx (µ, ν) = µ(x)ifor
Hw
F (x)
someix ∈ (0, 1],ithen theifunction x
iis convex.iAlso,iif Cx (µ, ν) = ν(x)ifor
Hw
F (x)
someix ∈ [0, 1),ithen theifunctioni 1−x
iisiconvex.

(iv) Considering Fiand wiare twoicontinuousifunctions oni[κ1 , κ2 ].iIf Fiis non-

negative (non-positive) oni[κ1 , κ2 ], thenithe functioniHw
F (x)iis increas-
ingi(decreasing)ion [0, 21 )iandiis decreasingi(increasing)ion ( 12 , 0].iAlso, Hw
F (x)ihas

airelative extremeipointiat x = 1
2
.iIf w ̸≡ 0,ithenicorresponding toianyia ∈
[κ1 , κ2 ] \ { κ1 +κ
2
2
}isatisfying

F(a) + F(κ1 + κ2 − a) = 0, i

thereiexists aicritical pointifor Hw

F (x).i

Proof. The mapping Hw

F (x) exhibits similar symmetry, convexity, and monotonicity

Properties to those of the mapping Mw

F (x) derived in proposition 6.

Theorem 3.3.1. Considering an h-convexifunctioniF : [κ1 , κ2 ] ⊆ R → Riandiw :

[κ1 , κ2 ] → R+ ∪ {0}.iThe mappingiHw
F : [0, 1] → Ridefined asiabove satisfies the fol-

lowing

(i) Theimapping Hw
F iis convexioni[0, 1].

(ii)

 Z Cx (µ,ν) Z κ2 
Cx (µ, ν) + Dx (µ, ν)
inf Hw
F (x) = Hw
F (0) =F w(a) da + w(a) da
x∈[0,1] 2 κ1 Dx (µ,ν)

(iii)
Z Cx (µ,ν) Z κ2
sup Hw
F (x) = Hw
F (1) = F(a)w(a) da + F(a)w(a) da.
x∈[0,1] κ1 Dx (µ,ν)

65
Proof. (i) Letir, s ≥ 0iwithir + s = 1iand x1 , x2 ∈ [0, 1]. then

Hw
F (rx1 + sx2 )
Z Cx (µ,ν)  

Cx (µ, ν) + Dx (µ, ν)

= F (rx1 + sx2 )a + 1 − (rx1 + sx2 ) w(a) da
κ1 2
Z κ2  

Cx (µ, ν) + Dx (µ, ν)

+ F (rx1 + sx2 )a + 1 − (rx1 + sx2 ) w(a) da
Dx (µ,ν) 2
Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)

= F r x1 a + (1 − x1 ) w(a) da
κ1 2
Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)

+ F s x2 a + (1 − x2 ) w(a) da
κ1 2
Z κ2  
Cx (µ, ν) + Dx (µ, ν)

+ F r x1 a + (1 − x1 ) w(a) da
Dx (µ,ν) 2
Z κ2  
Cx (µ, ν) + Dx (µ, ν)

+ F s x2 a + (1 − x2 ) w(a) da
Dx (µ,ν) 2
Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)
≤r F x1 a + (1 − x1 ) w(a) da
κ1 2
Z κ2   
Cx (µ, ν) + Dx (µ, ν)
+ F x1 a + (1 − x1 ) w(a) da
Dx (µ,ν) 2
Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)
+s F x2 a + (1 − x2 ) w(a) da
κ1 2
Z κ2   
Cx (µ, ν) + Dx (µ, ν)
+ F x2 a + (1 − x2 ) w(a) da = rHw w
F (x1 ) + sHF (x2 ).
Dx (µ,ν) 2

(ii) At x = 0 we have
Z Cx (µ,ν)  
w Cx (µ, ν) + Dx (µ, ν)

HF (0) = F (0)a + (1 − 0) w(a) da
κ1 2
Z κ2  
Cx (µ, ν) + Dx (µ, ν)

+ F (0)a + (1 − 0) w(a) da
Dx (µ,ν) 2
Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)
= F w(a) da
κ1 2
Z κ2  
Cx (µ, ν) + Dx (µ, ν)
+ F w(a) da
Dx (µ,ν) 2
 Z Cx (µ,ν) Z κ2 
Cx (µ, ν) + Dx (µ, ν)
=F w(a) da + w(a) da = inf Hw F (x).
2 κ1 Dx (µ,ν) x∈[0,1]

66
Now at x = 1 we have
Z Cx (µ,ν)  
Cx (µ, ν) + Dx (µ, ν)

Hw
F (1) = F (1)a + (1 − 1) w(a) da
κ1 2
Z κ2  
Cx (µ, ν) + Dx (µ, ν)

+ F (1)a + (1 − 1) w(a) da
Dx (µ,ν) 2
Z Cx (µ,ν) Z κ2
= F(a)w(a) da + F(a)w(a) da = sup Hw F (x).
κ1 Dx (µ,ν) x∈[0,1]

Hence proved.

67
Chapter 4

CONCLUSIONS AND FUTURE

RECOMMENDATION

This thesis has contributed to the understanding and analysis of convex functions par-
ticularly focusing on p-convex functions, h-convexifunctions, and stronglyireciprocallyi(p, h)-
convexifunctions forihigher-order cases. These generalizations are essential tools
inivarious disciplines ofipure andiappliedisciences.iThis study aims to examine these
generalized convex functions in connection with well-known inequalities, as a result,
we can deeply understand the behavior, properties, and practical significance.
In the introductory chapter, we gave a detailed review of the convex function
that covers the historical background, its applications in both pure and applied
sciences, and interpret its various properties, which set a strong base for further
study. Moreover, we discussed some well-known inequalities in connection with
convexity, as these are pivotal for scientific study. Further in Chapter 2, we fo-
cus onistrongly reciprocallyi(p, h)-convexifunctions forihigher-order, a new version
oficonvexifunction.iWe investigate some basic properties that enable us to explore

68
Hermite-Hadamardiinequality, Fejér inequality, and fractional integral inequalities
within this context. An important contribution of this thesis is the introduction of
new mappings discussed in Chapter 3. We began with introducing a novel mapping

F for h-convex functions where its dynamic properties led to the composition
Mw
of ainew generalizediFejér-typeiinequality andiHermite-Hadamardiinequality with
aideeper understanding. A second mapping Hw
F , along with several associated results

offers additional tools for further analysis of h-convexifunction. Through the dynamic
properties of these mappings, we can extend the current literature to new research
avenues, particularly in the study of inequality theory.

69
Bibliography

[1] C. Niculescu and L.-E. Persson, Convex functions and their applications, vol. 23.
Springer, 2006.

[2] J. E. N. Valdés, F. Rabossi, and A. D. Samaniego, “Convex functions: Ariadne’s

thread or Charlotte’s spiderweb?,” Advanced Mathematical Models & Applications,
vol. 5, no. 2, 2020.

[3] O. Holder, “Uber einen mittelwertssatz,” Nachr. Akad. Wiss. Gottingen Math.-
Phys. Kl., 1889.

[4] O. Stolz, Grundzüge der Differential-und Integralrechnung. BG Teubner, 1893.

[5] C. Hermite et al., “Sur deux limites d’une intégrale définie,” Mathesis, vol. 3,
no. 82, p. 1, 1883.

[6] J. Hadamard, “Essai sur l’étude des fonctions données par leur développement
de taylor; étude sur les propriétés des fonctions entières et en particulier d’une
fonction considérée par riemann...,” (No Title), 1893.

[7] J. L. W. V. Jensen, “Om konvekse funktioner og uligheder imellem middelvaerdier,”

Nyt tidsskrift for matematik, vol. 16, pp. 49–68, 1905.

70
 [8] J. L. W. V. Jensen, “Sur les fonctions convexes et les inégalités entre les valeurs
moyennes,” Acta mathematica, vol. 30, no. 1, pp. 175–193, 1906.

[9] C.-H. Tan, “Convex sets and convex functions,” Journal of Convex Analysis,
vol. 26, no. 3, pp. 837–856, 2019.

[10] B. Lafferriere, G. Lafferriere, and M. N. Nguyen, Introduction to Mathematical

Analysis I. Portland State University Library, 2022.

[11] J. Nachbar, “Concave and convex functions,” Lecture Notes for Economics,
vol. 4111, 2018.

[12] B. T. Nguyen and P. D. Khanh, “Lipschitz continuity of convex functions,” Applied

Mathematics & Optimization, vol. 84, no. 2, pp. 1623–1640, 2021.

[13] K. Zhang and J. Wan, “p-convex functions and their properties,” Pure Appl. Math,
vol. 23, no. 1, pp. 130–133, 2007.

[14] İ. İşcan, “Hermite-Hadamard type inequalities for harmonically convex functions,”

Hacettepe Journal of Mathematics and statistics, vol. 43, no. 6, pp. 935–942, 2014.

[15] S. Varošanec, “On h-convexity,” Journal of Mathematical Analysis and Applica-

tions, vol. 326, no. 1, pp. 303–311, 2007.

[16] U. S. Kirmaci, M. K. Bakula, M. E. Özdemir, and J. Pečarić, “Hadamard-type in-

equalities for s-convex functions,” Applied Mathematics and Computation, vol. 193,
no. 1, pp. 26–35, 2007.

[17] K. Murota and A. Tamura, “New characterizations of m-convex functions and

their applications to economic equilibrium models with indivisibilities,” Discrete
Applied Mathematics, vol. 131, no. 2, pp. 495–512, 2003.

71
[18] R. D. Nussbaum, “Convexity and log convexity for the spectral radius,” Linear
Algebra and its Applications, vol. 73, pp. 59–122, 1986.

[19] M. Eshaghi, S. S. Dragomir, and M. Rostamian Delavar, “An inequality related to

η -convex functions (ii),” International Journal of Nonlinear Analysis and Appli-
cations, vol. 6, no. 2, pp. 27–33, 2015.

[20] A. A. Shaikh, A. Iqbal, and C. K. Mondal, “Some results on φ-convex functions

and geodesic φ-convex functions,” arXiv preprint arXiv:1706.05841, 2017.

[21] B. Micherda and T. Rajba, “On some hermite-hadamard-fejér inequalities for (k,
h)-convex functions,” Math. Inequal. Appl, vol. 15, no. 4, pp. 931–940, 2012.

[22] J. E. N. Valdés, F. Rabossi, and A. D. Samaniego, “Convex functions: Ariadne’s

thread or charlotte’s spiderweb?,” Advanced Mathematical Models & Applications,
vol. 5, no. 2, 2020.

[23] L. Azócar, M. Bracamonte, and J. Medina, “Some inequalities of jensen type

and lazhar type for the class of harmonically and strongly reciprocally convex
functions,” Applied Mathematics, vol. 11, no. 4, pp. 1075–1080, 2017.

[24] L. Fejér, “Über die fourierreihen, ii,” Math. Naturwiss. Anz Ungar. Akad. Wiss,
vol. 24, no. 369.390, 1906.

[25] H. Chen and U. N. Katugampola, “Hermite–hadamard and hermite–hadamard–

fejér type inequalities for generalized fractional integrals,” Journal of Mathematical
Analysis and Applications, vol. 446, no. 2, pp. 1274–1291, 2017.

72
[26] M. R. Delavar and S. S. Dragomir, “Weighted trapezoidal inequalities related to
the area balance of a function with applications,” Applied Mathematics and Com-
putation, vol. 340, pp. 5–14, 2019.

[27] M. R. Delavar and S. Dragomir, “Two mappings in connection to fejér inequality

with applications,” Math. Inequal. Appl., vol. 21, no. 4, pp. 1111–1123, 2018.

[28] G. Farid and A. U. Rehman, “Generalization of the fejer-hadamard’s inequality

for convex function on coordinates,” Communications of the Korean Mathematical
Society, vol. 31, no. 1, pp. 53–64, 2016.

[29] I. Işcan, “Hermite-hadamard-fejér type inequalities for convex functions via frac-
tional integrals,” arXiv preprint arXiv:1404.7722, 2014.

[30] D. Kotrys, “Remarks on jensen, hermite-hadamard and fejer inequalities for

strongly convex stochastic processes,” Mathematica Aeterna, vol. 5, no. 1, p. 104,
2015.

[31] M. Rostamian Delavar, S. Mohammadi Aslani, and M. De La Sen, “Hermite-

hadamard-fejér inequality related to generalized convex functions via fractional
integrals,” Journal of Mathematics, vol. 2018, no. 1, p. 5864091, 2018.

[32] M. Rostamian Delavar, “On fejér’s inequality: generalizations and applications,”

Journal of Inequalities and Applications, vol. 2023, no. 1, p. 42, 2023.

[33] M. Z. Sarikaya and H. Budak, “On fejér type inequalities via local fractional inte-
grals,” Journal of Fractional Calculus and Applications, vol. 8, no. 1, pp. 59–77,
2017.

73
[34] E. Set, İ. İşcan, M. Z. Sarikaya, and M. E. Özdemir, “On new inequalities of
hermite–hadamard–fejér type for convex functions via fractional integrals,” Ap-
plied Mathematics and Computation, vol. 259, pp. 875–881, 2015.

[35] K.-L. Tseng, G.-S. Yang, and K.-C. Hsu, “Some inequalities for differentiable map-
pings and applications to fejér inequality and weighted trapezoidal formula,” Tai-
wanese journal of Mathematics, vol. 15, no. 4, pp. 1737–1747, 2011.

[36] H.-N. Shi and J. Zhang, “Some new judgement theorems of schur geometric and
schur harmonic convexities for a class of symmetric functions,” Journal of Inequal-
ities and Applications, vol. 2013, pp. 1–9, 2013.

[37] B. T. Polyak, “Existence theorems and convergence of minimizing sequences for

extremal problems with constraints,” in Doklady Akademii Nauk, vol. 166, pp. 287–
290, Russian Academy of Sciences, 1966.

[38] S. Maden, S. Turhan, and İşcan, “Hermite-hadamard inequality for strongly p-

convex functions,” June 2018. ResearchGate preprint.

[39] M. A. Noor, K. I. Noor, and S. Iftikhar, “Hermite-hadamard inequalities for har-

monic nonconvex functions,” MAGNT Res. Rep, vol. 4, no. 1, pp. 24–40, 2016.

[40] M. Bracamonte, J. Medina, and M. Vivas, “On some inequalities for strongly
reciprocally convex functions,” Extracta Mathematicae, vol. 33, no. 1, pp. 109–
122, 2018.

[41] H. Li, M. S. Saleem, I. Hussain, and M. Imran, “Strongly reciprocally p-convex

functions and some inequalities,” Journal of Mathematics, vol. 2020, no. 1,
p. 4957141, 2020.

74
[42] Z. B. Fang and R. Shi, “On the (p, h)-convex function and some integral inequal-
ities,” Journal of Inequalities and Applications, vol. 2014, pp. 1–16, 2014.

[43] G.-H. Lin and M. Fukushima, “Some exact penalty results for nonlinear programs
and mathematical programs with equilibrium constraints,” Journal of Optimiza-
tion Theory and Applications, vol. 118, pp. 67–80, 2003.

[44] H. Li, M. S. Saleem, I. Ahmed, and K. N. Aslam, “Hermite–Hadamard and Fejer-

type inequalities for strongly reciprocally (p, h)-convex functions of higher order,”
Journal of Inequalities and Applications, vol. 2023, no. 1, p. 57, 2023.

[45] G. Farid and B. Tariq, “Some integral inequalities for m-convex functions via
fractional integrals,” Journal of Inequalities and Special Functions, vol. 8, no. 1,
pp. 170–185, 2017.

[46] G. Farid and G. Abbas, “Generalizations of some hermite–hadamard–fejér type

inequalities for p-convex functions via generalized fractional integrals,” Journal of
Fractional Calculus and Applications, vol. 9, no. 2, pp. 56–76, 2018.

[47] B.-Y. Xi and F. Qi, “Some integral inequalities of hermite-hadamard type for
convex functions with applications to means,” Journal of function spaces, vol. 2012,
no. 1, p. 980438, 2012.

[48] M. A. Noor, K. I. Noor, and S. Iftikhar, “Integral inequalities for differentiable

p-harmonic convex functions,” Filomat, vol. 31, no. 20, pp. 6575–6584, 2017.

[49] E. Godunova and V. Levin, “Neravenstva dlja funkcii sirokogo klassa, soderza-
scego vypuklye, monotonnye i nekotorye drugie vidy funkii. vycislitel. mat. i,”
Fiz. Mezvuzov. Sb. Nauc. Trudov, MGPI, Moskva, vol. 9, pp. 138–142, 1985.

75
[50] W. W. Breckner, “Stetigkeitsaussagen für eine klasse verallgemeinerter konvexer
funktionen in topologischen linearen räumen,” Publications de l’Institut Mathéma-
tique (Beograd) (NS), vol. 23, no. 37, pp. 13–20, 1978.

[51] S. S. Dragomir, J. Pecaric, and L.-E. Persson, “Some inequalities of hadamard

type,” Soochow journal of mathematics, vol. 21, no. 3, pp. 335–341, 1995.

[52] S. S. Dragomir, D. M. Milošević, and J. Sándor, “On some refinements of

hadamard’s inequalities and applications,” Publikacije Elektrotehničkog Fakulteta.
Serija Matematika, pp. 3–10, 1993.

[53] K.-L. Tseng, S.-R. Hwang, and S. S. Dragomir, “Fejér-type inequalities (i),” Jour-
nal of Inequalities and Applications, vol. 2010, pp. 1–7, 2010.

[54] K.-L. Tseng, S.-R. Hwang, and S. S. Dragomir, “Fejér-type inequalities (ii),” Math-
ematica Slovaca, vol. 67, no. 1, pp. 109–120, 2017.

[55] U. S. Kirmaci, M. K. Bakula, M. E. Özdemir, and J. Pečarić, “Hadamard-type in-

equalities for s-convex functions,” Applied Mathematics and Computation, vol. 193,
no. 1, pp. 26–35, 2007.

[56] C. E. Pearce and A. Rubinov, “P-functions, quasi-convex functions, and hadamard-

type inequalities,” Journal of Mathematical Analysis and Applications, vol. 240,
no. 1, pp. 92–104, 1999.

[57] M. Z. Sarikaya, A. Saglam, and H. Yildirim, “On some hadamard-type inequal-

ities for h-convex functions,” Journal of Mathematical Inequalities, vol. 2, no. 3,
pp. 335–341, 2008.

76
[58] M. Z. Sarikaya, E. Set, and M. E. Ozdemir, “On some new inequalities of hadamard
type involving h-convex functions,” Acta Mathematica Universitatis Comenianae
(NS), vol. 79, no. 2, pp. 265–272, 2010.

77