MTH600(A) Final Year Project (2024)

Project's Title:

A Comparative Study on Hermite-Hadamard type Inequalities for Convex Functions with

Applications
A Report Submitted to: Department of Mathematics

Supervisor Name: Dr. Jamshed Nasir

Supervisor Email: jamshed@vu.edu.pk

Submitted By:

Group ID: S2407AFD71

Sr. No Student ID Student Name

1 BC230215232 SHAHER BANO (GR)
2 BC220404939 TOOBA MUMTAZ
3 BC220406887 AYESHA ATEEQ
4 BC220424113 FIZZA NOOR
5 BC220406501
6 BC220416017
7 BC220420150

1
Table of contents:

https://testbook.com/maths/convex-function

2
 Chapter 1
Fundamental concepts of Convex Sets and Convex Functions

Introduction:
The concept of convexity plays a vital role in various branches of mathematics, particularly in
optimization theory, functional analysis, and geometry. A convex set is a fundamental concept in
mathematics, defined as a set that contains all the points on the line segment connecting any two
points within the set. This property ensures that the set is "curved inward" and has no "dents" or
"holes".
A convex function is a function that preserves the convexity of the domain. In other words, if the
domain is a convex set, then the function is convex if it satisfies the convexity inequality.
Convex functions have numerous applications in optimization problems, economics, physics,
and engineering, among other fields.
Hermite-Hadamard's inequalities are a class of fundamental inequalities in mathematics that
provide a bound for the value of a convex function at the midpoint of an interval. These
inequalities have far-reaching implications in various fields, including optimization theory,
numerical analysis, and physics.
Means are an essential concept in mathematics, and different types of means have been
extensively studied. The most common types of means include:
 Arithmetic Mean (AM)
 Geometric Mean (GM)
 Harmonic Mean (HM)
 Quadratic Mean (QM)
 Power Mean (PM)
These means have various applications in mathematics, physics, engineering, and other fields.
Hermite-Hadamard type inequalities for convex functions have been extensively studied and
generalized in various directions. These inequalities provide a powerful tool for estimating the
value of a convex function at a midpoint, and have numerous applications in optimization
problems, numerical analysis, and physics.
In this project, we aim to conduct a comparative study of different versions of Hermite-
Hadamard type inequalities for convex functions, including their statements, proofs, and
applications. We will explore the historical development of these inequalities, their
interconnections, and their implications in various fields.
Through this project, we hope to provide a comprehensive understanding of Hermite-Hadamard
type inequalities and their significance in modern mathematics and science. We will also
investigate the applications of these inequalities in real-world problems, highlighting their
relevance and utility in addressing practical challenges.

3
1.1 Preliminaries
This chapter introducing fundamental concepts and definitions essential for understanding the subsequent
chapters. A review of the definition and properties of convex sets, including the concept of convex hulls
and convex combinations. An exploration of the definition and characteristics of convex functions,
including the concept of convex conjugates and subgradients. An, introduction to the classical Hermite-
Hadamard inequalities, which provide bounds on the value of a convex function over a closed interval.

1.2 Convex Set

A convex set is a collection of points in which the line AB connecting any two points A, B in the
set lies completely within the set. In other words, A subset S of En is considered to be convex if
any linear combination
θ x 1+ (1−θ ) x 2 ,(0≤ θ ≤ 1)
is also included in S for all pairs of x 1 , x 2 ∈ S .

Figure 1: Convex Set

1.2.1 Example1:
An interval of [a , b]⊂ R is a convex set. To see this, let c , d ∈[a , b] and assume, without loss of
generality, that c <d . Let λ ∈(0 , 1). Then
a ≤ c=( 1−λ ) c + λc < ( 1−λ ) c+ λd
¿ ( 1− λ ) d+ λd=d
≤b
1.2.2 Example 2:
n
In Rn the set H ≔{x ∈ R :a1 x 1+ …+an x n=c } is a convex set. For any particular choice of
constants a i it is hyperplane in Rn . It’s defining the equation is a generalization of the usual
equation of a plane in R3, namely the equation ax +by + cz+ d=0 .
to see that H is a convex set, let x(1) , x(2) ∈ H and define z ∈ R3 by z ≔ ( 1−λ ) x (1 )+ λx( 2). Then
n
z=∑ ai [ ( 1−λ ) x i(1 )+ λ x i(2 ) ]
i=1
n
¿ ∑ (1−λ)ai x i(1) + λ a i x i(2 )
i=1

4
 n n
¿ ( 1− λ ) ∑ ai xi +¿ λ ∑ a i x i(2 ) ¿
( 1)

i=1 i=1
¿ ( 1− λ ) c+ λc
¿c
Hence z ∈ H .

1.3 Non-Convex Set:

Non-convex sets are those that are not convex. A non-convex polygon is occasionally referred to
as a concave polygon, and also some sources use the phrase concave set to refer to a non-convex
set.

Figure 2: Non-Convex set

1.3.1 Extreme and Non-extreme Points of Convex Sets

Now, consider the below figure which represents the extreme and non-extreme points of convex
sets.

Figure 3: Extreme and non-extreme point of Convex set

It is noted that the points x1, x2, x3, x4 and x5 are extreme points of convex set S, whereas x6 and
x7 are the non-extreme points of the convex set.
Thus, extreme point or vertex of a convex set S is a point x that does not fall on any line segment
connecting any two unique points in S, such as x1 and x2.

5
1.4 Convex Sets Examples

Prove that C = {(x1, x2): 2x1 + 3x2 = 7} ⊂ R2 is a convex set.

1.4.1 Example 1:

Assume that X, Y ∈ C, where X = (x1, x2), Y = (y1, y2). The line segment connecting X and Y is
Solution:

the set.
From the definition of convex sets, we can write the following:
W = {W: W = θ X + (1 − θ) Y, 0 ≤ θ ≤ 1}
For some 0 ≤ θ ≤ 1, assume that W = (w1, w2) is the point of set W.
Hence, we can write
w1 = θx1 + (1 − θ) y1

As x, y ∈ C, we can write
w2 = θx2 + (1 − θ) y2

2x1 + 3x2 = 7, and

2y1 + 3y2 = 7
But, from the formula,
2w1 + 3w2 = 2[θx1 + (1 − θ) y1] + 3[θx2 + (1 − θ) y2]
Now, take the common terms outside, we get
= θ [2x1 + 3x2] + (1 − θ) [2y1 + 3y2]

Hence, W = (w1, w2) belongs to C Since W is any point of C, and X, Y ∈ C

= θ.7 + (1 − θ).7 = 7. [Since 2x1 + 3x2 = 7]

This can be written as [X: Y] ⊂ C.

Therefore, set C is convex.

Prove that the set B = {(x1, x2, x3): 2x1 − x2 + x3 ≤ 4} ⊂ R3 is a convex set.
1.4.2 Example 2:

Solution:
Assume that X = (x1, x2, x3) and Y = (y1, y2, y3) are the two points of B.
From the given conditions, we can write
2x1 − x2 + x3 ≤ 4
2y1 − y2 + y3 ≤ 4
Now, assume that W = (w1, w2, w3) is any point of [X, Y] such that 0 ≤ θ ≤ 1,
w1 = θx1 + (1 − θ) y1
w2 = θx2 + (1 − θ) y2
w3 = θx3 + (1 – θ) y3
Using the above equations, we can write
2w1 − w2 + w3 = θ (2x1 − x2 + x3) + (1 − θ) (2y1 − y2 + y3) ≤ 4θ + 4(1 − θ) = 4
Thus, W = (w1, w2, w3) is a point of S.
Therefore, X, Y belongs to B, and hence [ X ; Y ]⊂ B .
Hence, the set B is a convex set.

6
1.5 Types of Convex Sets:

1.5.1 Open Convex Set:

A convex set that does not include its boundary points is known as open convex set.
Example:
The open disk {(x , y )∨x 2 + y 2 <1}.

1.5.2 Closed Convex Set:

A convex set that includes its boundary points is known as closed convex set.
Example:
The closed disk {(x , y )∨x 2 + y 2 ≤ 1 }.

1.5.3 Convex Combinations

Let x 1 , x 2 ∈ V . Then the set of all convex combinations of x 1and x 2 is the set of points
{ ω λ ∈V : ω λ =( 1−λ ) x 1 + λ x 2 ,0 ≤ λ ≤ 1 }

1.5.4 Strictly Convex Set:

A convex set that contains no line segments.
Example:
The open disk {(x , y )∨x 2 + y 2 <1}is strictly convex

1.5.6 Convex Hull

The convex hull of a set ℂ is the intersection of all convex sets which contain the set ℂ. We
denote the convex hull by co ( C ) .

Figure 4: Formation of a Convex Hull

Example1:
Suppose that [a , b] and [c , d ] are two intervals of the real line with b< c so that the intervals are
disjoint. Then the convex hull of the set [a , b]∪ [c , d ] is just the interval [a , d ].

7
1.6 Closure of Convex set
The closure of a convex set is the smallest closed set that contains the convex set. In other words,
it's the set of all points that are either in the convex set or arbitrarily close to it.

Formally, the closure of a convex set C is denoted as cl(C) or C̄ , and it's defined as:
cl(C )={x∨there ∃a sequence {x n }∈C such that x n → x }

1.6.1 Key properties of the closure of a convex set:

 cl(C) is a closed set (by definition).

 cl(C) is convex (since the limit of a sequence of convex combinations is also a convex
combination).

cl(C) is the smallest closed set containing C (if K is a closed set and C ⊆ K, then cl(C) ⊆
 C ⊆ cl(C) (the original convex set is contained in its closure).

K).

8
1.6.2 Examples:
 The closure of an open interval (a, b) is the closed interval [a, b].
 The closure of a convex polygon is the polygon itself (since it's already closed).
 The closure of a convex set like the open unit disk (x ¿ ¿ 2+ y 2 <1) ¿ is the closed unit disk
( x 2+ y 2 ≤ 1 ).

Let A and B be two convex sets. We need to show that their union, A ∪ B, is also convex.
1.7 Union of Convex Sets

To prove this, let's take two points x and y in A ∪ B. We need to show that the line segment
between x and y, denoted as [x, y], is also in A ∪ B.
There are three possible cases:
a) Both x and y are in A.
b) Both x and y are in B.
c) One of x or y is in A, and the other is in B.

Therefore, [x, y] is in A ∪ B.
Case 1: If both x and y are in A, then [x, y] is in A (since A is convex).

Therefore, [x, y] is in A ∪ B.
Case 2: If both x and y are in B, then [x, y] is in B (since B is convex).

Case 3: Suppose x is in A and y is in B (or vice versa). Let z be a point on the line segment [x, y].
Then, z can be written as:

for some λ ∈ [0, 1].

z=λx +(1−λ) y

Since A is convex, the point λx +(1−λ)a is in A for any a in A. Similarly, since B is convex, the
point λb+(1− λ) y is in B for any b in B.
Let a=x∧b= y . Then, we get:
z=λx +(1−λ) x=x (which is in A)
or

Therefore, z is in A ∪ B.
z=λy +(1−λ) y= y (which is in B)

In all three cases, we have shown that [x, y] is in A ∪ B. Hence, A ∪ B is convex.

Note: This proof assumes that the convex sets A and B are defined in a real vector space. The
result holds for convex sets in any real vector space, including Euclidean spaces and infinite-
dimensional spaces.

1.8 Intersection of convex sets:

Let A and B be convex sets. We need to show that the intersection A ∩ B is also convex.
Let x , y ∈ A ∩ B , and let λ ∈[0 , 1]. We need to show that λx +(1−λ) y ∈ A ∩B .
Since x , y ∈ A and A is convex, we have:
λx +(1−λ) y ∈ A
Similarly, since x , y ∈ B and B is convex, we have:
λx + ( 1−λ ) y ∈ B
Therefore, λx +(1−λ) y ∈ A ∩B , and thus A ∩ B is convex.

9
 This proof can be extended to the intersection of any number of convex sets. Let { A i } be a
collection of convex sets. We need to show that ∩i Ai is convex.

Let x , y ∈∩i A i, and let λ ∈ [0, 1]. Then, x , y ∈ A i ∀ i. Since Ai is convex, we have:
λx +(1−λ) y ∈ A i
Therefore, λx +(1−λ) y ∈ ∩i A i , and thus ∩ i Ai is convex.
This proof shows that the intersection of any number of convex sets is always convex.

Separation point
Let S , T ⊂ Rn and let H be a hyperplane. Then H is said to separate S from T if S lies in one
closed half-space determined by H while T lies in the other closed half-space. In this case H is
called a separating hyperplane. If S and T lie in the open half-spaces, then H is said to strictly
separate S and T.
Extreme point:
polytope
A set which can be expressed as the intersection of a finite number of half-spaces is said to be a convex
polytope

A function f (x) is convex on an interval [a, b] if for any two points x 1and x 2 in [a, b] and any 𝜆
1.1 Convex Function

where 0 < 𝜆 < 1,

f [ λ x 1 + ( 1− λ ) x2 ] ≤ λf ( x1 ) +(1−λ) f (x 2)
If f ( x ) has a second derivative in [a, b], then a necessary and sufficient condition for it to be
convex on that interval is that the second derivative f '' ( x ) ≥ 0 ∀ x ∈[a , b].
Examples of convex functions [1];
 e ax
 −log (x¿)¿
a
 |a| , a ≥1
 x log ( x ) ¿

10
1.1.1 Proper Convex Function
A function f (x) is said to be proper convex if it satisfies the following two conditions:
 The function is convex, which means that for any two points x 1and x 2 in its domain and
any 𝜆 in the interval [0, 1], the following inequality holds:
f [ λ x 1 + ( 1− λ ) x2 ] ≤ λf ( x1 ) + ( 1−λ ) f ( x 2 )
 The function has no finite values of negative infinity, which means that there does not
exist any x in its domain such that
f ( x )=−∞
1.1.4

[1] Mitrinovic, D. S., & Vasic, P. M. (1970). Analytic inequalities (Vol. 61). Berlin: Springer.

11