Professional Documents
Culture Documents
Research Article
Hermite–Hadamard Type Inequalities via Generalized Harmonic
Exponential Convexity and Applications
Received 5 January 2021; Revised 23 January 2021; Accepted 31 January 2021; Published 12 February 2021
Copyright © 2021 Saad Ihsan Butt et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In this work, we introduce the idea and concept of m–polynomial p–harmonic exponential type convex functions. In addition, we
elaborate the newly introduced idea by examples and some interesting algebraic properties. As a result, several new integral
inequalities are established. Finally, we investigate some applications for means. The amazing techniques and wonderful ideas of
this work may excite and motivate for further activities and research in the different areas of science.
introduced a new class called n–polynomial harmonically If we put p = −1 and p = 1, then p–harmonic convex sets
convex function. Motivated and inspired by the ongoing and p–harmonic convex functions collapse to classical con-
activities and research in the convex analysis field, we found vex sets, harmonic convex sets, and harmonic convex func-
out that there exists a special class of function known as tions, respectively.
exponential convex function, and nowadays, a lot of people
working are in this field [12, 13]. Dragomir [14] introduced Definition 4 (see [17]). A function ψ : I ⊆ ð0,+∞Þ ⟶ ½0,+
the class of exponential type convexity. After Dragomir, ∞Þ is called m–polynomial harmonic exponential type con-
Awan et al. [15] studied and investigated a new class of expo- vex function, if
nentially convex functions. Kadakal and İşcan introduced a
new definition of exponential type convexity in [16].
Recently, Geo et al. [17] introduced n–polynomial harmonic ℘1 ℘2 1 m
ψ ≤ 〠 ðeκ − 1Þ j ψð℘1 Þ
exponential type convex functions. The fruitful benefits and κ℘2 + ð1 − κÞ℘1 m j=1
applications of exponential type convexity is used to manip- ð5Þ
ulate for statistical learning, information sciences, data min- 1 m j
+ 〠 e1−κ − 1 ψð℘2 Þ,
ing, stochastic optimization and sequential prediction [7, m j=1
18, 19] and the references therein. Before we start, we need
the following necessary known definitions and literature.
holds for every ℘1 , ℘2 ∈ I,m ∈ ℕ and κ ∈ ½0, 1.
2. Preliminaries
Motivated by the above results, literature, and ongoing
In this section, we recall some known concepts.
activities and research, we organise the paper in the following
way. Firstly, we will give the idea and its algebraic properties
Definition 1 (see [5]). Let ψ : I ⟶ ℝ be a real valued func-
of m–polynomial p–harmonic exponential type convex func-
tion. A function ψ is said to be convex, if
tions. Secondly, we will derive the new sort of (H–H) and
refinements of (H–H) type inequalities by using the newly
ψðκ℘1 + ð1 − κÞ℘2 Þ ≤ κψð℘1 Þ + ð1 − κÞψð℘2 Þ, ð1Þ introduced idea. Finally, we will give some applications for
means and conclusion.
holds for all ℘1 , ℘2 ∈ I and κ ∈ ½0, 1:
3. Generalized Exponential Type Convex
Definition 2 (see [20]). A function ψ : I ⊆ ð0,∞Þ ⟶ ℝ is said Functions and Its Properties
to be harmonic convex, if
We are going to introduce a new definition called m–polyno-
℘1 ℘2 mial p–harmonic exponential type convex function and will
ψ ≤ κψð℘1 Þ + ð1 − κÞψð℘2 Þ, ð2Þ
κ℘2 + ð1 − κÞ℘1 study some of their algebraic properties. Throughout the
paper, one thing gets in mind m represents finite ℤ+ , m–poly
holds for all ℘1 , ℘2 ∈ I and κ ∈ ½0, 1. p–harmonic exp convex function represents m–polynomial p
–harmonic exponential type convex function and (H–H)
For the harmonic convex function, İşcan [20] provided represents Hermite–Hadamard.
the Hermite–Hadamard type inequality.
Definition 5. A function ψ : I ⊆ ð0,+∞Þ ⟶ ½0,+∞Þ is called
Definition 3 (see [21]). A function ψ : I ⟶ ℝ is said to be p – m–poly p–harmonic exp convex, if
harmonic convex, if
" #1/p !
" #1/p ! p p
℘1 ℘2 1 m κ
p p
℘1 ℘2 ψ p p ≤ 〠 ð e − 1 Þ j ψ ð℘ 1 Þ
ψ p p ≤ κψð℘1 Þ + ð1 − κÞψð℘2 Þ, ð3Þ κ℘2 + ð1 − κÞ℘1 m j=1
κ℘2 + ð1 − κÞ℘1
1 m 1−κ j
+ 〠 e − 1 ψð℘2 Þ,
holds for all ℘1 , ℘2 ∈ I and κ ∈ ½0, 1. m j=1
Note that κ = 1/2 in (3), we get the following inequality: ð6Þ
" p
# !
p 1/p
2℘1 ℘2 ψð℘1 Þ + ψð℘2 Þ
ψ ≤ , ð4Þ holds for every ℘1 , ℘2 ∈ I,m ∈ ℕ and κ ∈ ½0, 1.
p p
℘1 + ℘2 2
Remark 6.
holds for all ℘1 , ℘2 ∈ I.
(i) Taking m = 1 in Definition 5, we obtain the following
The function ψ is called Jensen p–harmonic convex new definition about p –harmonically exp convex
function. function:
Journal of Function Spaces 3
Proof. The rest of the proof is clearly seen. Theorem 13. Let ψ, φ : ½℘1 , ℘2 ⟶ ℝ: If ψ and φ are two m–
poly p–harmonic exp convex functions, then
Proposition 8. Every p–harmonic convex function is m–poly
p–harmonic exp convex function. (i) ψ + φ is an m–poly p–harmonic exp convex function
4 Journal of Function Spaces
(ii) For c ∈ ℝðc ≥ 0Þ, cψ is an m–poly p–harmonic exp (iv) If p = −1 in Theorem 13, then as a result, we get the
convex function ψ + φ and cψ are m–poly exp convex functions
(v) If m = 1 and p = −1 in Theorem 13, then as a result,
Proof. we get Theorem 2.1 in [16]
Remark 16.
(ii) Let ψ be an m–pol p–harmonic exp convex, then
(i) In case of m = 1, we investigate the following new
inequality:
" p p
#1/p !
℘1 ℘2
ðcψÞ p
κ℘2 + ð1 − κÞ℘1
p " p p
#1/p !
" # ℘1 ℘2
1 m κ 1 m 1−κ j ðφ ∘ ψÞ p p ≤ ð e κ − 1 Þ ðφ ∘ ψ Þ ð ℘ 1 Þ
≤c j
〠 ð e − 1 Þ ψ ð℘ 1 Þ + 〠 e − 1 ψ ð℘ 2 Þ t℘2 + ð1 − t Þ℘1
m j=1 m j=1
+ e1−κ − 1 ðφ ∘ ψÞð℘2 Þ
1 m
1 m
j
= 〠 ðeκ − 1Þ j cψð℘1 Þ + 〠 e1−κ − 1 cψð℘2 Þ ð16Þ
m j=1 m j=1
1 m κ 1 m j
= 〠 ðe − 1Þ j ðcψÞð℘1 Þ + 〠 e1−κ − 1 ðcψÞð℘2 Þ
m j=1 m j=1 (ii) In case of p = 1, the above Theorem 15 collapses to
ð14Þ Theorem 3.3 in [17]
(iii) In case of m = p = 1, as a result, we obtain the follow-
which completes the proof. ing new inequality:
Remark 14.
℘1 ℘2
ðφ ∘ ψÞ ≤ ðeκ − 1Þðφ ∘ ψÞð℘1 Þ
(i) If m = 1 in Theorem 13, then as a result, we get the κ℘2 + ð1 − κÞ℘1 ð17Þ
ψ + φ and cψ are p–harmonic exp convex functions
+ e1−κ − 1 ðφ ∘ ψÞð℘2 Þ
(ii) If p = 1 in Theorem 13, then as a result, we get
Theorem 3.2 in [17]
(iii) If m = p = 1 in Theorem 13, then as a result, we get (iv) In case of p = −1, then, the above Theorem 15
the ψ + φ and cψ are harmonic exp convex functions collapses to the following new inequality:
Journal of Function Spaces 5
1/p !
Theorem 19. If ψ : ½℘1 , ℘2 ⟶ ℝ is an m–poly p–harmonic 2xp yp 1 m pffiffi j 1 m pffiffi j
ψ ≤ 〠 e − 1 ψðxÞ + 〠 e − 1 ψðyÞ:
exp convex then ψ is bounded on ½℘1 , ℘2 : xp + y p m j=1 m j=1
Proof. Let x ∈ ½℘1 , ℘2 and L = max fψð℘1 Þ, ψð℘2 Þg, then, ð23Þ
p p p
there exist ∃κ ∈ ½0, 1 such that x = ½ð℘1 ℘2 Þ / ðκ℘2 + ð1 − κÞ
p 1/p κ 1−κ Using the change of variables, we get
℘1 Þ : Thus, since e ≤ e and e ≤ e, we have
" p p
#1/p ! " #1/p ! ( " #1/p !
℘1 ℘2 1 m 1 m pffiffi j
p p p p
2℘1 ℘2 ℘1 ℘2
ψðx Þ = ψ ≤ 〠 ðeκ − 1Þ j ψð℘1 Þ ψ p p ≤ 〠 e−1 × ψ p p
p p
κ℘2 + ð1 − κÞ℘1 m j=1 ℘1 + ℘2 m j=1 κ℘2 + ð1 − κÞ℘1
" p p
#1/p !)
1 m 1−κ j 1 m j
+ψ p 1 2
℘℘
+ 〠 e − 1 ψð℘2 Þ ≤ 〠 eκ + e1−κ − 2 ⋅ L p
κ℘1 + ð1 − κÞ℘2
:
m j=1 m j=1
2L m ð24Þ
≤ 〠 ðe − 1Þ j = M:
m j=1
Integrating the above inequality with respect to κ on
ð20Þ ½0, 1, we obtain
6 Journal of Function Spaces
" p p
#1/p ! p p ð ℘2 (ii) In case of m = 1 and p = −1, then as a result, we
m 2℘1 ℘2 p℘ ℘ ψðν Þ
m pffiffi j ψ ≤ p 1 2p dν, obtain Theorem 3.1 in [16]
℘1 ν
p p
2∑ j=1 e − 1 ℘1 + ℘2 ℘2 − ℘1 p+1
(iii) In case of m = 1 and p = 1, then as a result, we obtain
ð25Þ Corollary 1in [17]
Theorem 27. Let ψ : I = ½℘1 , ℘2 ⊆ ℝ \ f0g ⟶ ℝ be differen- Corollary 29. Under the assumptions of Theorem 27 with
q
tiable function on the I ∘ of I. If ψ ′ ∈ L½℘1 , ℘2 and jψ ′ j is an p = 1, we have the following new result:
m-poly p–harmonic exp convex function on I, q ≥ 1, then
ð ℘2
ψ ð ℘ Þ + ψ ð℘ Þ ℘ ℘ ψ ð Þ
℘ ℘ ð℘ − ℘ Þ
x
1 2
− 1 2
≤ 1 2 2 1
℘ ℘ ℘p − ℘p
p p ð℘
dx
ψ ð ℘ Þ + ψ ð℘ Þ ℘ 2 ψðx Þ
2 ℘2 − ℘1 ℘1 x2
2
p℘
1 2
− p 1 2
dx
≤ 1 2 2 1
h
2 p
℘2 − ℘1 ℘1 x 1+p
2p 1−ð1/qÞ
q
q i1/q
G1 ′ G2′
ψ ′ ð℘1 Þ
+ G′3
ψ ′ ð℘2 Þ
,
h
q
q i1/q
1−ð1/qÞ
G1 G2
ψ ′ ð℘1 Þ
+ G3
ψ ′ ð℘2 Þ
,
ð38Þ
ð34Þ
where
where ð1 ð
1 1 j1 − 2κj∑ j=1 ðeκ − 1Þ
m j
j1 − 2κj
G′1 = dκ, G2
′ = dκ,
A2κ m 0 A2κ
ð1 ð 0
1 1 ∣1 − 2κ ∣ ∑ j=1 ðeκ − 1Þ
m j
∣1 − 2κ ∣ ð m j
G1 = dκ, G 2 = dκ, 1 1 j1 − 2κj∑ j=1 e1−κ − 1
0 Aκ
p+1 m 0 A1+p
κ G′3 = dκ:
m 0 A2κ
ð m j
1 1 ∣1 − 2κ ∣ ∑ j=1 e1−κ − 1 ð39Þ
G3 = dκ:
m 0 A1+p
κ
ð35Þ Theorem 30. Let ψ : I = ½℘1 , ℘2 ⊆ ℝ \ f0g ⟶ ℝ be differ-
entiable function on the I ∘ of I. If ψ ′ ∈ L½℘1 , ℘2 and
q
jψ ′ j is an m–poly p–harmonic exp convex function on I,
Proof. Using Lemma 25, properties of modulus, power mean
r, q ≥ 1, ð1/rÞ + ð1/qÞ ≥ 1, then,
inequality, and m–poly p–harmonic exp convexity of the
q
jψ ′ j , we have
ψ ð ℘ Þ + ψ ð℘ Þ p p ð℘
℘ 2 ψðx Þ
℘ ℘ ℘ p − ℘ p
p℘
2 1
1 2
− 1 2
dx
≤
1 2
ψð℘ Þ + ψð℘ Þ p p ð
p p ð
2 p
℘2 − ℘1 a
p
x 1+p
2p
p℘ ℘ ℘2 ψðxÞ
℘1 ℘2 ℘2 − ℘1 1 j1 − 2κj
1
2
2
− p 1 2p
℘2 − ℘1 ℘1 x1+p
dx
≤
2p 0 Aκ
p+1 h
q
q i1/q
p p ð 1 1−ð1/qÞ × G1/r 4 G5
ψ ′ ð ℘ Þ
+ G 6
ψ ′ ð ℘ Þ
,
℘℘
℘ ℘ ℘ − ℘1 j1 − 2κj 1 2
ψ ′ 1 2
dκ ≤ 1 2 2 dκ
Aκ 2p 0 Aκ
p+1
ð 1
j1 − 2κj
℘℘
q 1/q p
℘ ℘ ℘ − ℘1
p ð 1
j1 − 2κj
1−ð1/qÞ ð40Þ
p+1
ψ ′ 1 2
dκ ≤ 1 2 2 dκ
0 A κ
A κ 2p 0 A p+1
κ
0 h
i 11/q
ð 1 j1 − 2κj ð1/mÞ∑m ðeκ − 1Þ j
ψ ′ ð℘ Þ
q + ð1/mÞ∑m e1−κ − 1 j
ψ ′ ð℘ Þ
q where
j=1 1 j=1 2
×@ dκA
0 A1+p
κ
ð1 ð1
℘ ℘ ℘ − ℘1
p p ð 1
j1 − 2κj
1−ð1/qÞ 1 ∑mj=1 ðeκ − 1Þ j
≤ 1 2 2
2p p+1
dκ G4 = j1 − 2κjr dκ, G5 = ð1+pÞq
dκ,
0 Aκ
!1/q 0 m 0 Aκ
ð ð1 m j
1 1 ∣1 − 2κ ∣ ∑ j=1 ðeκ − 1Þ
m j
q
dκ + 1
j1 − 2κj∑ j=1 e1−κ − 1
j ð41Þ
× ψ ′ ð ℘ Þ
ψ ′ ð℘ Þ
q dκ ð1
m 0 A1+p
κ
1
m 0 A1+p
κ
2
1 ∑mj=1 e1−κ − 1
p
℘1 ℘2 ℘2 − ℘1
p h
i1/q
G6 = ð1+pÞq
dκ:
≤ G1
1−ð1/qÞ
G2
ψ ′ ð℘1 Þ
+ G3
ψ ′ ð℘2 Þ
q q
, m 0 Aκ
2p
ð36Þ
Proof. Using Lemma 25, properties of modulus, Hölder’s
inequality, and m–poly p–harmonic exp convexity of the
q
which completes the proof. jψ ′ j , we have
8 Journal of Function Spaces
p p ð
p p
" " #1/p ! " #1/p ! #
ψð ℘ Þ + ψ ð ℘ Þ p℘1 ℘2 ℘2 ψðxÞ
℘1 ℘2 ℘2 − ℘1
1 p p p p
1 2
3℘1 ℘2 3℘1 ℘2
− dx
≤
2 ℘2 − ℘1 ℘1 x1+p
p p
2p
ψð℘1 Þ + 3ψ + 3ψ + ψ ð℘ Þ
8 p
℘1 + 2℘2
p p
2℘1 + ℘2
p 2
ð1
p p
j1 − 2κj
℘℘
℘ ℘ ℘ − ℘ p p ð
p p
ψ ′ 1 2
dκ ≤ 1 2 2 p℘ ℘ ℘2 ψðxÞ
℘1 ℘2 ℘2 − ℘1
1
p+1
0 Aκ Aκ 2p − p 1 2p dx
≤
ð 1 1/r ð 1
!1/q ℘2 − ℘1 ℘1 x1+p
p
1
℘1 ℘2
q
j1 − 2κ j r
dκ ψ ′ dκ h
q i1/q
ð1+pÞq
Aκ
1−ð1/qÞ
B4
ψ ′ ð℘1 Þ
+ B5
ψ ′ ð℘2 Þ
q
0 0 Aκ B1
p p ð 1 1/r
℘ ℘ ℘ − ℘1 h
≤ 1 2 2 j1 − 2κjr dκ 1−ð1/qÞ
q
q i1/q
2p 0 + B2 B6
ψ ′ ð℘1 Þ
+ B7
ψ ′ ð℘2 Þ
ð1 " # !1/q
q 1 m 1−κ j
q h
×
1 1 m κ j
′
〠 ðe − 1Þ ψ ð℘1 Þ + 〠 e − 1 ψ ð℘2 Þ dκ
′
1−ð1/qÞ
q
q i1/q
0 Aκ
ð1+pÞq m
j=1 m j=1 + B3 B8
ψ ′ ð℘1 Þ
+ B9
ψ ′ ð℘2 Þ
,
p h
℘ ℘ ℘ − ℘1
p
q i1/q
= 1 2 2 G1/r4 G 5
ψ ′ ð ℘1 Þ
q
+ G 6
ψ ′ ð ℘2 Þ
, ð46Þ
2p
ð42Þ where
ð 1/3 ð 2/3
jκ − ð1/8Þj jκ − ð1/2Þj
which completes the proof. B1 = dκ, B2 = dκ,
0 Ap+1
κ 1/3 Ap+1
κ
Corollary 31. Under the assumptions of Theorem 30 with ð 1/3
jκ − ð1/8Þj∑ j=1 ðeκ − 1Þ j
m
1
p = −1, we have the following new result: B4 = dκ,
m 0 Ap+1
κ
ð 1/3 m j
ψ ð ℘ Þ + ψ ð℘ Þ ð ℘2
1 jκ − ð1/8Þj∑ j=1 e1−κ − 1
1
B5 = dκ,
1 2
− ψðxÞdx
m Ap+1
2 ℘2 − ℘1 ℘1
0 κ
ð 1 1/r m ð 2/3
jκ − ð1/2Þj∑ j=1 ðeκ − 1Þ j
m
ð℘ − ℘1 Þ 1 1
≤ 2 j1 − 2κjr dκ 〠 ðe − 2 Þ j ð43Þ B6 = dκ,
2 0 m m 1/3 Ap+1
κ
j=1
h m j
q
q i1/q ð
1 2/3 jκ − ð1/2Þj∑ j=1 e1−κ − 1
ψ ′ ð℘1 Þ
+
ψ ′ ð℘2 Þ
: B7 = dκ,
m 1/3 Ap+1
κ
ð1
jκ − ð7/8Þj∑ j=1 ðeκ − 1Þ j
m
1
Corollary 32. Under the assumptions of Theorem 30 with B8 = dκ,
p = 1, we have the following new result:
m
2/3 Ap+1
κ
ð m j
1 1 jκ − ð7/8Þj∑ j=1 e1−κ − 1
ð
B9 = dκ:
ψ ð ℘ Þ + ψ ð℘ Þ ℘1 ℘2 ℘2 ψðxÞ
℘1 ℘2 ð℘2 − ℘1 Þ m 2/3 Ap+1
κ
1 2
− dx
≤
2 ℘2 − ℘1 ℘1 x2
2 ð47Þ
h i
q
q 1/q
G4 ′ G5′
ψ ′ ð℘1 Þ
+ G6′
ψ ′ ð℘2 Þ
1/r
, Proof. Using Lemma 26, properties of modulus, power
mean inequality, and m–poly p-harmonic exp convexity
q
ð44Þ of the jψ ′ j , we have
" " #1/p ! " #1/p ! # p p ð
1 p℘1 ℘2 ℘2 ψðxÞ
p p p p
3℘1 ℘2 3℘1 ℘2
ψð℘1 Þ + 3ψ + 3ψ + ψ ð ℘ 2 Þ − dx
where
8 p
℘1 + 2℘2
p p
2℘1 + ℘2
p
℘2 − ℘1 ℘1 x1+p
p p
p p ð 1/3
ð 2/3
℘ ℘ ℘ − ℘1 jκ − ð1/8Þj
℘1 ℘2
jκ − ð1/2Þj
℘1 ℘2
≤ 1 2 2 ×
ψ ′ A
dκ +
ψ ′ A
dκ
p 0 A1+p κ 1/3 A1+p κ
ð1 ð m ð1
κ
# p " ð
κ
1 1 ∑ j=1 ðeκ − 1Þ
j 1−ð1/qÞ
jκ − ð7/8Þj
℘1 ℘2
p 1/3
℘1 ℘2 ℘2 − ℘1 jκ − ð1/8Þj
G′4 = j1 − 2κj dκ, G5′ =
r
dκ, +
A1+p
ψ ′ A
dκ ≤ p
×
A1+p
dκ
2/3 κ κ 0 κ
0 m 0 A2q
κ ð 1/3
1/q ð 2/3 1−ð1/qÞ
ð45Þ jκ − ð1/8Þj
℘1 ℘2
q jκ − ð1/2Þj
ð m j 1+p
ψ ′ A
dκ + 1+p
dκ
1 1 ∑ j=1 e1−κ − 1 0 A κ κ 1/3 A κ
ð 2/3
1/q ð 1 1−ð1/qÞ
G6′ = dκ:
jκ − ð1/2Þj
ψ ′
℘1 ℘2
q
dκ +
jκ − ð7/8Þj
dκ
m 0 A2q
κ 1/3 A 1+p
κ
A κ
2/3 A 1+p
κ
ð 1
1/q # p p
"ð 1−ð1/qÞ
jκ − ð7/8Þj
℘1 ℘2
q ℘1 ℘2 ℘2 − ℘1 1/3
jκ − ð1/8Þj
1+p
ψ ′
dκ ≤ × 1+p
dκ
2/3 A κ
A κ p 0 A κ
0 h
i 11/q
Theorem 33. Let ψ : I = ½℘1 , ℘2 ⊆ ℝ \ f0g ⟶ ℝ be differ- ð 1/3 jκ − ð1/8Þj ð1/mÞ∑m ðeκ − 1Þ j
ψ ′ ð℘ Þ
q + ð1/mÞ∑m e1−κ − 1 j
ψ ′ ð℘ Þ
q
j=1 1 j=1 2
× @ dκA
entiable function on the I ∘ of I. If ψ ′ ∈ L½℘1 , ℘2 and 0 A1+p
κ
q
jψ ′ j is an m–poly p–harmonic exp convex function on I, +
ð 2/3
jκ − ð1/2Þj
dκ
1−ð1/qÞ
0 h
i 11/q where
ð 2/3 jκ − ð1/2Þj ð1/mÞ∑m ðeκ − 1Þ j
ψ ′ ð℘ Þ
q + ð1/mÞ∑m e1−κ − 1 j
ψ ′ ð℘ Þ
q
j=1 1 j=1 2
×@ dκA
A1+p ð 1/3 κ
1/3 κ
1 ∑m
j=1 ðe − 1Þ
j
ð 1
jκ − ð7/8Þj
1−ð1/qÞ B10 = ð1+pÞq
dκ,
+
A1+p
dκ m Aκ 0
2/3 κ
h
i 11/q #
m j
0
ð 1 jκ − ð7/8Þj ð1/mÞ∑m ðeκ − 1Þ j
ψ ′ ð℘ Þ
q + ð 1∕ mÞ∑m e1−κ − 1 j
ψ ′ ð℘ Þ
q ð
×@
j=1 1 j=1 2
dκA 1 1/3 ∑ j=1 e1−κ − 1
2/3 A1+pκ B11 = ð1+pÞq
dκ,
p p
" ð 1/3 1−ð1/qÞ ð m 0 Aκ
1 1/3 ∣κ − ð1/8Þ ∣ ∑ j=1 ðeκ − 1Þ
m j
℘ ℘ ℘ − ℘1 jκ − ð1/8Þj
≤ 1 2 2 × ×
1+p
dκ 1+p ð
1 2/3 ∑ j=1 ðeκ − 1Þ
p 0 A κ
m 0 A κ m j
j !1/q
q 1
ð 1/3
∣κ − ð1/8Þ ∣ ∑mj=1 e1−κ − 1
q B12 = dκ,
ψ ′ ð℘1 Þ
dκ +
m 0 A1+p
ψ ′ ð℘ Þ
dκ
2 m 1/3 Aðκ1+pÞq
κ
ð51Þ
m j
ð 2/3 1−ð1/qÞ ð 2/3
jκ − ð1/2Þj 1 ∣κ − ð1/2Þ ∣ ∑ j=1 ðeκ − 1Þ j
m
ð
+ 1+p
dκ ×
A1+p
ψ ′ ð℘ Þ
q dκ
1 1 2/3 ∑ j=1 e1−κ − 1
1/3 Aκ m 1/3
!1/q
κ
!1−ð1/qÞ
B13 = ð1+pÞq
dκ,
ð m j ð m 1/3
+
1 2/3 ∣κ − ð1/2Þ ∣ ∑ j=1 e1−κ − 1
ψ ′ ℘
q
dκ +
1 1 jκ − ð7/8Þj Aκ
ð 2 Þ dt
m 1/3 A1+p
κ
m 2/3 A1+p ð1 κ
∑m
j=1 ðe − 1Þ
j
t
ð1 κ ð1 1−κ j 1
B14 =
m
1 ∣κ − ð 7/8Þ ∣ ∑ ð e − 1 Þ j
1 ∣κ − ð7/8Þ ∣ ∑m j=1 e −1
×
j=1
q
ψ ′ ð℘1 Þ dκ + ð1+pÞq
dκ,
m 2/3 Aκ 1+p m 2/3 Aκ 1+p
m
Aκ 2/3
!1/q # p p
m j
h
q i1/q
q
ψ ′ ð℘2 Þ
dκ
℘1 ℘2 ℘2 − ℘1 1−ð1/qÞ
B4
ψ ′ ð℘1 Þ
+ B5
ψ ′ ð℘2 Þ
q ð
=
p
B1
1 1 ∑ j=1 e1−κ − 1
h
q
q i1/q h
q
q i1/q
B15 = ð1+pÞq
dκ:
+ B2
1−ð1/qÞ
B6
ψ ′ ð℘1 Þ
+ B7
ψ ′ ð℘2 Þ
+ B3
1−ð1/qÞ
B8
ψ ′ ð℘1 Þ
+ B9
ψ ′ ð℘2 Þ
, m 2/3 Aκ
ð48Þ Proof. Using Lemma 26, properties of modulus, Hölder’s
inequality, and m–poly p–harmonic exp convexity of the
which completes the proof. q
jψ ′ j , we have
Corollary 34. Under the assumptions of Theorem 33 with
"
1
" p p
#1/p ! " p p
#1/p ! #
3℘1 ℘2 3℘1 ℘2
p = −1 and m = 1, we have the following new result:
ψð℘1 Þ + 3ψ + 3ψ + ψð℘2 Þ
8 p
℘1 + 2℘2
p p
2℘1 + ℘2
p
p p ð℘
p p ð 1/3
1
ψð℘ Þ + 3ψ 2℘1 + ℘2 + 3ψ ℘1 + 2℘2 + ψð℘ Þ p℘ ℘ ψðxÞ
℘ ℘ ℘ − ℘
κ − 1
2
1 2 2 1
− p 1 2p dx ≤ ×
8 1
3 3 2 ℘2 − ℘1 ℘1 x1+p
p 0
8
ð ℘2
ð 2/3
1 ℘ ℘
1
1 ℘ ℘
1
1+p ψ ′ 1 2
dκ +
κ −
ψ ′ 1 2
dκ
− ψðxÞdx
≤ ð℘2 − ℘1 Þ Aκ Aκ 1/3
2
A1+p Aκ
℘2 − ℘1 ℘1 ð1
κ
p p
h
7
1 ℘ ℘
℘1 ℘2 ℘2 − ℘ 1
17
q
q i1/q +
κ −
8
A1+p
ψ ′ 1
Aκ
2
dκ ≤
0:0069
ψ ′ ð℘1 Þ
+ 0:036
ψ ′ ð℘2 Þ
2/3
(ð
κ
p
576 1/3
r 1/r ð 1/3
!1/q
1
℘1 ℘2
q
κ − 1
dκ
0:183 h
q
q i1/q 17
8
ð1+pÞq
ψ ′
Aκ
dκ
+ ψ ′ ð℘1 Þ
+
ψ ′ ð℘2 Þ
+ 0 0 Aκ
360
576 ð 2/3
r 1/r ð 2/3
!1/q
h
q
q i1/q +
κ − 1
dκ
1
ψ
℘ ℘
q
′ 1 2
dκ
′
′
0:036 ψ ð℘1 Þ + 0:0069 ψ ð℘2 Þ
: ð49Þ 1/3
2
1/3 Aκ
ð1+pÞq
Aκ
ð 1
r 1/r ð 1
!1/q )
1
℘1 ℘2
q
+
κ − 7
dκ ψ ′ dκ
8
ð1+pÞq
Aκ
Theorem 35. Let ψ : I = ½℘1 , ℘2 ⊆ ℝ \ f0g ⟶ ℝ be differ- 2/3
p (ð
2/3 Aκ
p 1/3
r 1/r ð 1/3
℘ ℘ ℘ − ℘1
entiable function on the I ∘ of I. If ψ ′ ∈ L½℘1 , ℘2 and ≤ 1 2 2 ×
κ − 1
dκ ×
1
q p
8
ð1+pÞq
jψ ′ j is an m–poly p–harmonic exp convex function on I,
0 0 Aκ
" # !1/q
r, q ≥ 1 and ð1/rÞ + ð1/qÞ ≥ 1 then 1 m κ
q 1 m j
q
〠 ðe − 1Þ j
ψ ′ ð℘1 Þ
+ 〠 e1−κ − 1
ψ ′ ð℘2 Þ
dκ
m j=1 m j=1
" " #1/p ! " #1/p ! # ð 2/3
r 1/r ð 2/3 "
1 p p
3℘1 ℘2
p p
3℘1 ℘2
1
1 1 m κ
q
+
κ −
dκ × 〠 ðe − 1Þ j
ψ ′ ð℘1 Þ
ψð℘1 Þ + 3ψ + 3ψ + ψ ð ℘ Þ
2
ð 1+p Þq m j=1
8 p
℘1 + 2℘2
p p
2℘1 + ℘2
p 2 1/3 1/3 Aκ
( # !1/q ð
r 1/r
p p ð
1 m j
q 1
p℘ ℘ ℘2 ψðxÞ
℘1 ℘2 ℘2 − ℘1
κ − 7
dκ
p p 1/r
3r+1 + 5r+1 + 〠 e1−κ − 1
ψ ′ ð℘2 Þ
dκ +
− p 1 2p dx
≤ × m j=1
8
2/3
℘2 − ℘1 ℘ 1 x 1+p p 24 ðr + 1Þ
r+1
" # !1/q )
1/r ð1
q
q 1/q 1 1 m κ j
q 1 m
q
2 × 〠 ð e − 1 Þ ψ ′ ð℘1 Þ
+ 〠 e1−κ − 1 j
ψ ′ ð℘2 Þ
dκ
B10
ψ ′ ð℘1 Þ
+ B11
ψ ′ ð℘2 Þ
+ r+1 2/3 Aκ
ð1+pÞq m m j=1
6 ðr + 1Þ p p
j=1
( 1/r ð
r+1 1/r
q 1/q 1 1/3 ∑ j=1 ðeκ − 1Þ
m
℘ ℘ ℘ − ℘1 3 +5
r+1 r+1 j
q
q
3 + 5r+1 = 1 2 2 × ψ ′ ð℘1 Þ
dκ
B12 ψ ′ ð℘1 Þ + B13 ψ ′ ð℘2 Þ
+ p 24 ðr + 1Þ
r+1 m 0 Aκ
ð 1+p Þq
) 24 ðr + 1Þ
r+1
ð m j !1/q 1/r
1 1/3 ∑ j=1 e1−κ − 1
q 2
q
q 1/q + ψ ′ ð℘2 Þ
dκ +
B14
f ′ ðaÞ
dt + B15
ψ ′ ð℘ Þ
ð1+pÞq 6r+1 ðr + 1Þ
, 2
m 0 Aκ
ð 2/3 m κ ð 2/3 m 1−κ j !1/q
1 ∑ j=1 ðe − 1Þ
j
∑ j=1 e − 1
×
φ ′ ð℘ Þ
q dκ + 1
ψ ′ ð℘ Þ
q dκ
1 2
ð50Þ m 1/3 Aðκ Þq 1+p m 1/3 Aκ
ð1+p Þq
10 Journal of Function Spaces
ð53Þ pffiffiffi
Proof. Taking ψðνÞ = 1/ ν for ν > 0 in Theorem 21, then,
inequality (60) is easily captured.
6. Applications
Proposition 39. Let 0 < ℘1 < ℘2 and p ≥ 1. Then, we get the
In this section, we recall the following special means of two
following inequality:
positive number ℘1 , ℘2 with ℘1 < ℘2 :
(1) The arithmetic mean m p p p℘p1 ℘p2 ℘2 − ℘1
pffiffi j H ℘1 , ℘2 ≤ p
℘ 2 − ℘ 1 L ð℘ 1 , ℘ 2 Þ
p
℘1 + ℘2 2∑m
j=1 e−1
A = Að℘1 , ℘2 Þ = ð54Þ ð61Þ
2 p p 2 m
≤A ℘1 , ℘2 〠 ½e − 2 j :
m j=1
(2) The geometric mean
pffiffiffiffiffiffiffiffiffiffi Proof. Taking ψðνÞ = νp for ν > 0 in Theorem 21, then,
G = Gð℘1 , ℘2 Þ = ℘1 ℘2 ð55Þ inequality (61) is easily captured.
℘2 − ℘1
L = Lð℘1 , ℘2 Þ = ð57Þ Proof. Taking ψðνÞ = ν2 for ν > 0 in Theorem 21, then,
ln ℘2 − ln ℘1 inequality (62) is easily captured.
Journal of Function Spaces 11
Proposition 41. Let 0 < ℘1 < ℘2 and p ≥ 1. Then, we get the function have some nice properties. We proved that our
following inequality: new introduced class is very larger with respect to the known
class of functions, like m–polynomial convex and m–polyno-
p p ð ℘2
m p℘1 ℘2 −ln x mial harmonically convex. A new version of Hermite–Hada-
m pffiffi j ln Gð℘1 , ℘2 Þ ≤ p p dx mard type inequality and an integral identity for the
2∑ j=1 e − 1 ℘2 − ℘1 ℘1 x
p+1
differentiable function are obtained. It is high time to find
p p 2 m the applications of these inequalities along with efficient
≤ ln H p ℘1 , ℘2 〠 ½ e − 2 j : numerical methods. We believe that our new class of func-
m j=1
tions will have a very deep research in this fascinating field
ð63Þ of inequalities and also in pure and applied sciences. The
interesting techniques and wonderful ideas of this paper
Proof. Taking ψðνÞ = −ln ν for ν > 0 in Theorem 21, then, can be extended on the coordinates along with fractional cal-
inequality (63) is easily captured. culus. In the future, our goal is that we will continue our
research work in this direction furthermore.
Proposition 42. Let 0 < ℘1 < ℘2 . Then, we get the following
inequality: Data Availability
p p ð ℘2 Data will be provided on request to the first author.
m H ð℘ ,℘ Þ p℘1 ℘2 ex
m pffiffi j e 1 2 ≤ p p dx
2∑ j=1 e − 1 ℘2 − ℘1 ℘1 x
p+1
ð64Þ Conflicts of Interest
℘1 2 m
℘2
≤ Aðe , e Þ 〠 ½e − 2 j : The authors declare that there are no conflicts of interest
m j=1
associated with this publication.
Proof. Taking ψðνÞ = eν for ν > 0 in Theorem 21, then,
inequality (64) is easily captured.
Acknowledgments
The authors received financial support from the Taif Univer-
Proposition 43. Let 0 < ℘1 < ℘2 . Then, we get the following sity Researches Supporting Project number (TURSP-
inequality: 2020/031), Taif University, Taif, Saudi Arabia
p p ð
2 m p℘ ℘ ℘2 sin x References
Aðsin ℘1 , sin ℘2 Þ 〠 ½e − 2 j ≤ p 1 2p dx
m j=1 ℘2 − ℘1 ℘1 xp+1
m [1] S. S. Dragomir and V. Pearce, Selected Topics on Hermite-
≤ m pffiffi j sin H p ð℘1 , ℘2 Þ: Hadamard Inequalities and Applications, Victoria, RGMIA
2∑ j=1 e − 1 Monographs, 2000.
ð65Þ [2] K. Mehrez and P. Agarwal, “New Hermite–Hadamard type
integral inequalities for convex functions and their applica-
tions,” Journal of Computational and Applied Mathematics,
Proof. Taking ψðνÞ = sin ð−νÞ for ν ∈ ð0, ðπ/2ÞÞ in Theorem
vol. 350, pp. 274–285, 2019.
21, then, inequality (65) is easily captured.
[3] D. S. Mitrinović, J. E. Pečarić, and A. M. Fink, Classical and
new inequalities in analysis. Mathematics and its Applications
Remark 44. The above discussed means, namely, arithmetic, (East European Series), vol. 61, Kluwer Academic Publishers
geometric, harmonic, and logarithmic are well known in lit- Group, Dordrecht, 1993.
erature because these means have remarkable applications [4] Y. Khurshid, M. Adil Khan, and Y.-M. Chu, “Conformable
in machine learning, probability, statistics, and numerical integral inequalities of the Hermite-Hadamard type in terms
approximation [24]. But, in the future, we will try to find of GG- and GA-convexities,” Journal of Function Spaces,
the applications of the He Chengtian mean (also called as vol. 2019, Article ID 6926107, 8 pages, 2019.
He Chengtian average), which was introduced by the first [5] C. P. Niculescu and L. E. Persson, Convex Functions and Their
time a famous ancient Chinese mathematician He Chengtian Applications, Springer, New York, 2006.
[25]. The He Chengtian average was extended to solve non- [6] S. Varošanec, “On h–convexity,” Journal of Mathematical
linear oscillators and it is called as He’s max–min approach Analysis and Applications, vol. 326, pp. 303–311, 2007.
(also called as He’s max–min method), which was further [7] T. Toplu, M. Kadakal, and İ. İşcan, “On n-polynomial convex-
developed into a frequency–amplitude formulation for non- ity and some related inequalities,” AIMS Mathematics, vol. 5,
linear oscillators [26, 27]. no. 2, pp. 1304–1318, 2020.
[8] H. N. Shi and J. Zhang, “Some new judgement theorems of
7. Conclusion Schur geometric and schur harmonic convexities for a class
of symmetric function,” Journal of Inequalities and Applica-
We have introduced and investigated some algebraic proper- tions, vol. 2013, Article ID 527, 2013.
ties of a new class of functions, namely, m–poly p–harmonic [9] G. D. Anderson, M. K. Vamanamurthy, and M. Vuorinen,
exp convex. We showed that our new introduced class of “Generalized convexity and inequalities,” Journal of
12 Journal of Function Spaces
Mathematical Analysis and Applications, vol. 335, no. 2, [26] J. H. He, “Max-min approach to nonlinear oscillators,” Inter-
pp. 1294–1308, 2007. national Journal of Nonlinear Sciences and Numerical Simula-
[10] M. A. Noor, K. I. Noor, and S. Iftikhar, “Harmite–Hadamard tion, vol. 9, no. 2, pp. 207–210, 2008.
inequalities for harmonic nonconvex function,” MAGNT [27] J. H. He, “An improved amplitude-frequency formulation for
Research Report, vol. 4, no. 1, pp. 24–40, 2016. nonlinear oscillators,” International Journal of Nonlinear Sci-
[11] M. U. Awan, N. Akhtar, S. Iftikhar, M. A. Noor, and Y.- ences and Numerical Simulation, vol. 9, no. 2, pp. 211-212,
M. Chu, “New Hermite–Hadamard type inequalities for n- 2008.
polynomial harmonically convex functions,” Journal of
Inequalities and Applications, vol. 2020, Article ID 125, 2020.
[12] S. I. Butt, A. Kashuri, M. Tariq, J. Nasir, A. Aslam, and W. Gao,
“n–polynomial exponential type p–convex function with some
related inequalities and their applications,” Heliyon, vol. 6,
article e05420, 2020.
[13] S. I. Butt, A. Kashuri, M. Tariq, J. Nasir, A. Aslam, and W. Gao,
“Hermite–Hadamard-type inequalities via n-polynomial
exponential-type convexity and their applications,” Advances
in Difference Equations, vol. 2020, Article ID 508, 2020.
[14] S. S. Dragomir and I. Gomm, “Some Hermite–Hadamard's
inequality functions whose exponentials are convex,” Babes
Bolyai Math., vol. 60, pp. 527–534, 2015.
[15] M. U. Awan, N. Akhtar, S. Iftikhar, M. A. Noor, and Y. -M.
Chu, “Hermite–Hadamard type inequalities for exponentially
convex functions,” Applied Mathematics & Information Sci-
ences, vol. 12, no. 2, pp. 405–409, 2018.
[16] M. Kadakal and İ. İşcan, “Exponential type convexity and
some related inequalities,” Journal of Inequalities and Applica-
tions, vol. 2020, Article ID 82, 2020.
[17] W. Geo, A. Kashuri, S. I. Butt, M. Tariq, A. Aslam, and
M. Nadeem, “New inequalities via n–polynomial harmoniaclly
exponential type convex functions,” AIMS Mathematics, vol. 5,
no. 6, pp. 6856–6873, 2020.
[18] G. Alirezaei and R. Mahar, “On exponentially concave func-
tions and their impact in information theory,” in 2018 Infor-
mation Theory and Applications Workshop (ITA), pp. 1–10,
San Diego, CA, USA, 2018.
[19] S. Pal and T. K. L. Wong, “Exponentially concave functions
and a new information geometry,” The Annals of Probability,
vol. 46, no. 2, pp. 1070–1113, 2018.
[20] İ. İşcan, “Hermite-Hadamard type inequalities for harmoni-
cally convex functions,” Hacettepe Journal of Mathematics,
vol. 43, no. 6, pp. 935–942, 2014.
[21] M. A. Noor, K. I. Noor, and S. Iftikhar, “Integral inequalities
for diffrential p–harmonic convex function,” Filomet, vol. 31,
no. 20, pp. 6575–6584, 2017.
[22] M. A. Noor, K. I. Noor, M. U. Awan, and S. Costache, “Some
integral inequalities for harmonically h-convex functions,”
UPB Scientific Bulletin, Series A, vol. 77, no. 1, pp. 5–16, 2015.
[23] M. A. Noor, K. I. Noor, and S. Iftikhar, “Newton inequalities
for p-harmonic convex function,” Honam Mathematical Jour-
nal, vol. 40, no. 2, pp. 239–250, 2018.
[24] R. J. Dalpatadu, The Arithmetic–Geometric–Harmonic Mean,
JSM Math. Stat., 2014.
[25] J. H. He, “He Chengtian's inequality and its applications,”
Applied Mathematics and Computation, vol. 151, no. 3,
pp. 887–891, 2004.