Optimization

Techniques
Dr. Muhammad Saad
MATHEMATICAL
CONCEPTS
Lecture 01
Functions of Several Variables
Functions with one independent variable, as 𝑓(𝑥) = 𝑥𝑒 −𝑥 ,
or equations in two variables, such as 𝑥 2 + 𝑦 2 = 4,
describe curves in ℝ2 . If we now add a third variable to
the picture and consider functions of two independent
variables, as 𝑓(𝑥, 𝑦) = 𝑥 2 + 2𝑦 2 , or equations in three
variables, as 𝑥 2 + 𝑦 2 + 2𝑧 2 = 4, we get such functions and
equations describe surfaces in ℝ3 .
In general, functions of two variables are written in the
explicit form
𝑧 = 𝑓(𝑥, 𝑦)
or in the implicate form
𝐹(𝑥, 𝑦, 𝑧) = 0.

Muhammad Saad
Functions of Several Variables
A function 𝑧 = 𝑓(𝑥, 𝑦) assigns to each point (𝑥, 𝑦) in a set
D (called the domain) in ℝ2 a unique real number 𝑧 in as
subset of ℝ (called the range). It is notated by 𝑓: ℝ2 → ℝ.

Muhammad Saad
Functions of Several Variables

Muhammad Saad
 Functions of Several Variables
Level curves
Functions of two variables are represented by surfaces in
ℝ3 . However, functions can be represented in another
way, which is used to make topographic maps
Widely spaced
contours: slow
Closely spaced changes in
contours: rapid elevation
changes in
elevation

Muhammad Saad
Functions of Several Variables
Level curves
Consider a surface defined by the function 𝑧 = 𝑓(𝑥, 𝑦). Now
imagine stepping onto the surface and walking along a
path on which your elevation has the constant value 𝑧
= 𝑧0 . When the contour curve is projected onto the 𝑥𝑦-
plane, the result is the curve 𝑓(𝑥, 𝑦) = 𝑧0 . This curve in the
𝑥𝑦-plane is called a level curve.

Muhammad Saad
Functions of Several Variables
Level curves

Muhammad Saad
Functions of Several Variables
However, it is useful to take a brief look at functions of
more than two variables. Two such examples are
𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 − 2𝑥𝑦 + 𝑦 2 + 3𝑦𝑧 − 𝑧 2 + 4𝑥 − 2𝑦 + 3𝑥 − 6
and
𝑔 𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 = 𝑥12 − 4𝑥1 𝑥2 + 𝑥34 sin 𝑥1 𝑥4 − 3𝑥1 + 5𝑥2 cos 𝑥3 .
In the first function, (𝑥, 𝑦, 𝑧) represents a point ℝ3 , and the
function 𝑓 maps each point in space to a fourth quantity
in ℝ. In the second function, (𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 ) can represent a
point in ℝ. The function might map a point in ℝ4 to a fifth
quantity in ℝ. So, 𝑓: ℝ3 → ℝ and 𝑓: ℝ4 → ℝ. Also, we can us
the vector notation to write the second function as
𝑔 𝐱 = 𝑥12 − 4𝑥1 𝑥2 + 𝑥34 sin 𝑥1 𝑥4 − 3𝑥1 + 5𝑥2 cos 𝑥3 ,
where 𝐱 is the vector (𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 ) in ℝ4 .
Muhammad Saad
Partial Derivatives
The derivative of a function of one variable, 𝑦 = 𝑓(𝑥),
measures the rate of change of 𝑦 with respect to 𝑥, and it
gives slopes of tangent lines. The analogous idea for
functions of several variables presents a new twist:
Derivatives may be defined with respect to any of the
independent variables. The resulting derivatives are
called partial derivatives; they still represent rates of
change, and they are associated with slopes of tangents.
So, much of what you have learned about derivatives
applies to functions of several variables.

Muhammad Saad
Partial Derivatives
Notation
For a function 𝑓(𝑥1 , 𝑥2 , ⋯ , 𝑥𝑛 ), the partial derivative of 𝑓
with respect to 𝑥𝑖 is written as:
𝜕𝑓
or 𝑓𝑥𝑖 or 𝐷𝑥𝑖 𝑓
𝜕𝑥𝑖
The partial derivatives evaluated at a point (𝑎1 , 𝑎2 , ⋯ , 𝑎𝑛 )
are denoted in any of the following ways:
𝜕𝑓 𝜕𝑓
𝑎1 , 𝑎2 , ⋯ , 𝑎𝑛 = ቤ = 𝑓𝑥𝑖 (𝑎1 , 𝑎2 , ⋯ , 𝑎𝑛 )
𝜕𝑥𝑖 𝜕𝑥𝑖 (𝑎 ,𝑎 ,⋯,𝑎 )
1 2 𝑛

Muhammad Saad
Partial Derivatives

Muhammad Saad
Partial Derivatives
Calculating
All the rules and results for ordinary derivatives can be
used to compute partial derivatives. Specifically, to
compute 𝑓𝑥 𝑥, 𝑦, 𝑧 we treat 𝑦 and 𝑧 as constants and take
an ordinary derivative with respect to 𝑥. Similarly, to
compute 𝑓𝑦 and 𝑓𝑧 we treat the non-considered variables
as constants and differentiate with respect to the
considered variable.

Muhammad Saad
Partial Derivatives
Ex 01. Find the partial derivatives of the function:
𝐟 = 𝐱 𝟐 𝐬𝐢𝐧 𝐲 + 𝐲 𝐥𝐧 𝐳 − 𝐳 𝟒 𝐞𝐱
Answer.
𝜕𝑓
𝑓𝑥 = = 2𝑥 sin 𝑦 + 0 − 𝑧 4 𝑒 𝑥
𝜕x
𝜕𝑓
𝑓𝑦 = = x 2 cos 𝑦 + 1 ⋅ ln 𝑧 − 0
𝜕y
𝜕𝑓 1
𝑓𝑧 = = 0 + y ⋅ − 4𝑧 3 𝑒 𝑥
𝜕z z

Muhammad Saad
Partial Derivatives
Higher-Order Partial Derivatives
Just as we have higher-order derivatives of functions of
one variable, we also have higher-order partial
derivatives. For example, given a function 𝑓 and its
partial derivative 𝑓𝑥 , we can take the derivative of 𝑓𝑥 with
respect to 𝑥 or with respect to 𝑦.

Muhammad Saad
Partial Derivatives
Higher-Order Partial Derivatives
𝑓 = 𝑓(𝑥, 𝑦, 𝑧)
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒙 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒙 𝜕2 𝑓
𝑓 𝑓𝑥 = 𝑓𝑥𝑥 =
𝜕𝑥 𝜕𝑥 2

𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒚 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒚 𝜕2 𝑓
𝑓 𝑓𝑦 = 𝑓𝑦𝑦 =
𝜕𝑦 𝜕𝑦 2

𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒛 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒛 𝜕2 𝑓
𝑓 𝑓𝑧 = 𝑓𝑧𝑧 =
𝜕𝑧 𝜕𝑧 2

𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒙 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒚 𝜕2 𝑓
𝑓 𝑓𝑥 = 𝑓𝑥𝑦 =
𝜕𝑥 𝜕𝑦𝜕𝑥
Muhammad Saad
Partial Derivatives
Higher-Order Partial Derivatives
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒚 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒙 𝜕2 𝑓
𝑓 𝑓𝑦 = 𝑓𝑦𝑥 =
𝜕𝑦 𝜕𝑥𝜕𝑦
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒙 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒛 𝜕2 𝑓
𝑓 𝑓𝑥 = 𝑓𝑥𝑧 =
𝜕𝑥 𝜕𝑧𝜕𝑥
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒛 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒙 𝜕2 𝑓
𝑓 𝑓𝑧 = 𝑓𝑧𝑥 =
𝜕𝑧 𝜕𝑥𝜕𝑧
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒚 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒛 𝜕2 𝑓
𝑓 𝑓𝑦 = 𝑓𝑦𝑧 =
𝜕𝑦 𝜕𝑧𝜕𝑦
𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑤.𝑟.𝑡𝑜 𝒛 𝜕𝑓 𝑤.𝑟.𝑡𝑜 𝒚 𝜕2 𝑓
𝑓 𝑓𝑧 = 𝑓𝑧𝑦 =
𝜕𝑧 𝜕𝑦𝜕𝑧
Muhammad Saad
Partial Derivatives
Ex. 02 Find the second partial derivatives of the
function:
𝒘 = 𝒙 𝐬𝐢𝐧 𝒚𝒛 + 𝒛𝒆𝟐𝒚+𝟑𝒙
Answer.
1st derivatives
wx = sin yz + 3ze2y+3x
wy = xz cos yz + 2ze2y+3x
wz = xy cos yz + e2y+3x
2nd derivatives
wxx = 9ze2y+3x
wyy = −xz 2 sin yz + 4ze2y+3x
wzz = −xy 2 sin yz
Muhammad Saad
Partial Derivatives
Ex 02. Find the second partial derivatives of the
function:
𝒘 = 𝒙 𝐬𝐢𝐧 𝒚𝒛 + 𝒛𝒆𝟐𝒚+𝟑𝒙
Answer.
1st derivatives
wx = sin yz + 3ze2y+3x
wy = xz cos yz + 2ze2y+3x
wz = xy cos yz + e2y+3x
2nd derivatives
wxy = wyx = z cos yz + 6e2y+3x
wxz = wzx = y cos yz + 3e2y+3x
wyz = wzy = −xyz sin yz + 2e2y+3x
Muhammad Saad
The Gradient Vector
Let 𝑓: ℝ𝑛 → ℝ be differentiable at the point (𝑥1 ,𝑥2 , ⋯ ,𝑥𝑛 ).
The gradient of 𝑓 at (𝑥1 ,𝑥2 , ⋯ ,𝑥𝑛 ) is the vector-valued
function
∇𝑓 𝑥1 , ⋯ , 𝑥𝑛 = 𝑓𝑥1 𝑥1 , ⋯ , 𝑥𝑛 , 𝑓𝑥2 𝑥1 , ⋯ , 𝑥𝑛 , ⋯ , 𝑓𝑥𝑛 𝑥1 , ⋯ , 𝑥𝑛
𝑓𝑥1 𝑥1 , ⋯ , 𝑥𝑛
𝑓 𝑥 , ⋯ , 𝑥𝑛
= 𝑥2 1
⋮
𝑓𝑥3 𝑥1 , ⋯ , 𝑥𝑛

Muhammad Saad
The Gradient Vector
Ex. 03 Evaluate the gradient for 𝛟 = 𝒙𝟐 + 𝐬𝐢𝐧 𝒛 ⋅ 𝐥𝐧 𝒚 at
𝝅
the point (𝟐,𝟏, ).
𝟐
Answer.

𝜕𝜙 𝜕𝜙 𝜕𝜙
∇𝜙 x, y, z = 𝑖+ 𝑗+ 𝑘
𝜕𝑥 𝜕𝑦 𝜕𝑧
sin 𝑧
= 2𝑥 𝑖 + 𝑗 + cos 𝑧 ln 𝑦 𝑘
𝑦
𝜋
∇𝜙(2,1, ) = 4𝑖 + 𝑗
2

Muhammad Saad
The Hessian Matrix
Given a function 𝑓: ℝ𝑛 → ℝ, we say that 𝑓 is twice
differentiable, and write the derivative of ∇𝑓 as the
matrix
𝑓𝑥1 𝑥1 𝑓𝑥1 𝑥2 ⋯ 𝑓𝑥1 𝑥𝑛
2 𝑓𝑥2 𝑥1 𝑓𝑥2 𝑥2 ⋯ 𝑓𝑥2 𝑥𝑛
𝐷 𝑓 𝑥1 , 𝑥2 , ⋯ , 𝑥𝑛 =
⋮ ⋮ ⋱ ⋮
𝑓𝑥𝑛 𝑥1 𝑓𝑥𝑛 𝑥2 ⋯ 𝑓𝑥𝑛 𝑥𝑛
The matrix 𝐷2 𝑓(𝐱) is called the Hessian matrix of 𝑓 at 𝐱,
and is often also denoted 𝑭(𝐱).

Muhammad Saad
The Hessian Matrix
Ex. 04 Evaluate the Hessian matrix of the function
𝒇 𝒙, 𝒚, 𝒛 = 𝒙𝟐 𝒆𝒚 − sin 𝑧
𝝅
at the point 𝑷 𝟑,𝟎,
𝟐

Answer.
2𝑒 𝑦 2𝑥𝑒 𝑦 0
𝐷2 𝑓 𝑥, 𝑦, 𝑧 = 2𝑥𝑒 𝑦 𝑥2𝑒𝑦 0
0 0 sin 𝑧
Hence,
𝜋 2 6 0
𝐷2 𝑓 3,0, = 6 9 0
2
0 0 1

Muhammad Saad
The Hessian Matrix
Ex. 05 Compute the gradient and Hessian matrix at
the point (𝟎,𝟏,𝝅) of the function:
𝒇 𝒙,𝒚,𝒛 = 𝒆−𝒙 𝐬𝐢𝐧(𝒚𝒛)
Answer.
−𝑒 −𝑥 sin 𝑦𝑧
∇𝑓 𝑥, 𝑦, 𝑧 = 𝑒 −𝑥 𝑧 cos 𝑦𝑧
𝑒 −𝑥 𝑦 cos 𝑦𝑧
𝐷 2 𝑓 𝑥, 𝑦, 𝑧
𝑒 −𝑥 sin 𝑦𝑧 −𝑒 −𝑥 𝑧 cos 𝑦𝑧 −𝑒 −𝑥 𝑦 cos 𝑦𝑧
= −𝑒 −𝑥 𝑧 cos 𝑦𝑧 −𝑒 −𝑥 𝑧 2 sin 𝑦𝑧 −𝑒 −𝑥 𝑧𝑦 sin 𝑦𝑧 + 𝑒 −𝑥 cos 𝑦𝑧
−𝑒 −𝑥 𝑦 cos 𝑦𝑧 𝑒 −𝑥 𝑧𝑦 sin 𝑦𝑧 + 𝑒 −𝑥 cos 𝑦𝑧 −𝑒 −𝑥 𝑦 2 sin 𝑦𝑧

Muhammad Saad
The Hessian Matrix
Ex. 05 Compute the gradient and Hessian matrix at
the point (𝟎,𝟏,𝝅) of the function:
𝒇 𝒙,𝒚,𝒛 = 𝒆−𝒙 𝐬𝐢𝐧(𝒚𝒛)
Answer.
Hence,
0
∇𝑓 0,1, 𝜋 = −𝜋
−1
0 𝜋 1
𝐹 0,1, 𝜋 = 𝐷 2 𝑓 0,1, 𝜋 = 𝜋 0 −1
1 −1 0

Muhammad Saad
Directional Derivative
Partial derivatives tell us a lot about the rate of change of
a function with respect to its variables. For example, 𝑓𝑥 of
the function 𝑓(𝑥,𝑦) is the rate of changing 𝑓 when one
move in the positive 𝑥 direction; that is the vector 1,0 .
Also, 𝑓𝑦 measures the change with respect to moving in
the direction 0,1 . However, they are not enough to
measure the rate of change if you walk in an infinite
number of directions from that point and find a different
rate of change in every direction. The derivative of 𝑓 in
the direction of 𝑢 is called the directional derivative of
𝑓 at some point (𝑎,𝑏) in the direction of 𝑢.

Muhammad Saad
Directional Derivative

Muhammad Saad
Directional Derivative
The directional derivative of a function 𝑓: ℝ𝑛 → ℝ at the
point (𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ) in the direction 𝑢 ∈ ℝ𝑛 is given by
𝑢
𝐷𝑢 𝑓 𝑎1 , 𝑎2 , ⋯ , 𝑎𝑛 = ∇𝑓 ⋅
|𝑢|
where | 𝑢 | is the norm (modulus) of 𝑢. If 𝑢 is a unit
vector, then simply
𝐷𝑢 𝑓 𝑎1 , 𝑎2 , ⋯ , 𝑎𝑛 = ∇𝑓 ⋅ 𝑢

Muhammad Saad
Directional Derivative
Ex. 06 Find the directional derivative of
𝝓 = 𝒙𝟐 𝒚𝒛 + 𝟒𝒙𝒛𝟐
𝟐
at the point (𝟏, − 𝟏,𝟏) in the direction 𝐮 = −𝟏 .
−𝟐
Answer.
2𝑥𝑦𝑧 + 4𝑧 2 2
∇𝜙 = 𝑥2𝑧 ⇒ ∇𝜙(1, −1,1) = 1
𝑥 2 𝑦 + 8𝑥𝑧 7
𝟐/𝟑
𝑢
= −𝟏/𝟑
𝑢 −𝟐/𝟑
4 1 14 11
𝐷𝑢 𝜙 1, −1,1 = − − =−
3 3 3 3
Muhammad Saad
Directional Derivative
Using properties of the dot product, we have
𝑢 1
𝐷𝑢 𝑓 = ∇𝑓 ⋅ = | ∇𝑓 𝑢 cos 𝜃) = | ∇𝑓| cos 𝜃 ,
𝑢 𝑢
where 𝜃 is the angle between 𝑢 and ∇𝑓. It follows that
the maximum value of directional derivative is | ∇𝑓 | when
cos 𝜃 = 1, that when 𝑢 and ∇𝑓 at the corresponding point
in the same direction. Here, the actual rate of increase is
𝐷𝑢 𝑓 = ||∇𝑓||
Similarly, 𝑓 has its greatest rate of decrease when ∇𝑓 and
𝑢 are in opposite directions. The actual rate of decrease is
𝐷𝑢 𝑓 = −||∇𝑓||

Muhammad Saad
Directional Derivative
Let 𝑓: ℝ𝑛 → ℝ be differentiable at (𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ) such that
∇ 𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ≠ 0.
1- 𝑓 has its maximum rate of increase at (𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ) in
the direction of the gradient ∇𝑓(𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ). The rate of
increase in this direction is ||∇𝑓(𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 )||.
2- 𝑓 has its maximum rate of decrease at (𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ) in
the direction of the gradient −∇𝑓(𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ). The rate of
decrease in this direction is −||∇𝑓(𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 )||.
3- The directional derivative is zero in any direction
orthogonal to ∇𝑓(𝑎1 ,𝑎2 , ⋯ ,𝑎𝑛 ).

Muhammad Saad
Tangent Plane
Let 𝐹(𝑥,𝑦,𝑧) be differentiable at the point 𝑃0 (𝑎,𝑏,𝑐) with
∇𝐹(𝑎,𝑏,𝑐) ≠ 0. The plane tangent to the surface 𝐹(𝑥,𝑦,𝑧) = 0
at 𝑃0 , called the tangent plane, is the plane passing
through 𝑃0 orthogonal to ∇𝐹(𝑎,𝑏,𝑐). An equation of the
tangent plane is
𝐹𝑥 𝑎, 𝑏, 𝑐 𝑥 − 𝑎 + 𝐹𝑦 𝑎, 𝑏, 𝑐 𝑦 − 𝑏 + 𝐹𝑧 𝑎, 𝑏, 𝑐 𝑧 − 𝑐 = 0.

Muhammad Saad
The Gradient and Level Curves
We have seen that in any direction orthogonal to the
gradient ∇𝑓(𝑎,𝑏), the function 𝑓 does not change at (𝑎,𝑏).
Recall, the curve 𝑓(𝑥,𝑦) = 𝑧0 , where 𝑧0 is a constant, is a
level curve, on which function values are constant. So, we
conclude that the gradient ∇𝑓(𝑎,𝑏) is orthogonal to the line
tangent to the level curve through (𝑎,𝑏).

Muhammad Saad
Positive Definity of Matrices
A symmetric matrix 𝐴 of order 𝑛 is called:
1- positive definite if 𝑥 𝑇 𝐴𝑥 > 0, for every nonzero vector
𝑥 ∈ ℝ𝑛 .
2- positive semidefinite if 𝑥 𝑇 𝐴𝑥 ≥ 0, for every nonzero
vector 𝑥 ∈ ℝ𝑛 .
3- negative definite if 𝑥 𝑇 𝐴𝑥 < 0, for every nonzero
vector 𝑥 ∈ ℝ𝑛 .
4- negative semidefinite if 𝑥 𝑇 𝐴𝑥 ≤ 0, for every nonzero
vector 𝑥 ∈ ℝ𝑛 .
5- Indefinite if it is not neither positive nor negative
semidefinite

Muhammad Saad
Positive Definity of Matrices
leading principal minors of a matrix
For a square matrix 𝑄 of odder 𝑛, there are always
leading principal minors Δ1 , Δ2 , ⋯ , Δ𝑛 as:

Muhammad Saad
Positive Definity of Matrices
Theorem: Let 𝐴 be a symmetric matrix of order 𝑛. Then
we have:
• 𝐴 is positive definite if and only if det(Δ𝑘 ) > 0, for all
leading principal minors Δ𝑘 of 𝐴.
• 𝐴 is positive semidefinite if and only ifdet(Δ𝑘 ) ≥ 0, for all
leading principal minors Δ𝑘 of 𝐴.
• 𝐴 is negative definite if and only if −1 k det(Δ𝑘 ) > 0, for
all leading principal minors Δ𝑘 of 𝐴.
• 𝐴 is negative definite if and only if −1 k det(Δ𝑘 ) ≥ 0, for
all leading principal minors Δ𝑘 of 𝐴.

Muhammad Saad
Positive Definity of Matrices
Ex. 07 Discuss the positivity of the following
matrices:
𝟑 𝟐
𝐀=
𝟎 𝟏
Answer.
It is not symmetric, then 𝐴 is not positive definite

Muhammad Saad
Positive Definity of Matrices
Ex. 07 Discuss the positivity of the following
matrices:
𝟐 𝟑
𝐀 =
𝟑 𝟓
Answer.
2 3
det(Δ1 ) = 2, det Δ2 = = 1, then the matrix is positive
𝟑 5
definite.

Muhammad Saad
Positive Definity of Matrices
Ex. 07 Discuss the positivity of the following
matrices:
𝟏 −𝟑 𝟐
𝐀 = −𝟑 𝟗 𝟏
𝟐 𝟏 𝟎
Answer.
det Δ1 = 1
1 −3
det Δ2 = =0
−3 9
𝟏 −𝟑 𝟕
det Δ2 = −𝟑 𝟗 𝟏 = −36
𝟕 𝟏 −𝟐
Thus, the matrix is 𝐴 indefinite.

Muhammad Saad
Positive Definity of Matrices
Ex. 07 Discuss the positivity of the following
matrices:
−𝟐 𝟏 𝟏
𝐀 = 𝟏 −𝟐 −𝟏
𝟏 −𝟏 −𝟐
Answer.
det Δ1 = −2
−2 1
det Δ2 = =3
1 −2
−2 1 1
det Δ2 = 1 −2 −1 = −6
1 −1 −2
Thus, the matrix is 𝐴 negative definit.

Muhammad Saad
Remember…
Study First!
(Keep Moving Forward)