Professional Documents
Culture Documents
A function 𝑓 defined on 𝐷 (the 𝐝𝐨𝐦𝐚𝐢𝐧 of 𝑓) is a rule that assigns , to each element (𝑥, 𝑦) of 𝐷, a
unique real number 𝑓(𝑥, 𝑦). x and 𝑦 are the independent variables. If we let 𝑧 = 𝑓(𝑥, 𝑦), then 𝑧 is
the dependent variable. The set { 𝑓(𝑥, 𝑦) : (𝑥, 𝑦) ∈ 𝐷} is the range of the function 𝑓.
The graph of a function 𝑓(𝑥, 𝑦) defined for (𝑥, 𝑦) ∈ 𝐷 is the set of points ( 𝑥, 𝑦, 𝑓(𝑥, 𝑦 ) ) in the
three-dimensional space with three mutually perpendicular axes : 𝑥-axis, 𝑦-axis and 𝑧–axis.
We call this graph a surface.
Graph of 𝒛 = 𝒇(𝒙, 𝒚)
(a surface)
Dr NG Wee Seng
Page 2
AY2018-19 Semester 1 MA1511 Engineering Calculus
Level Curves
Geometrically, for any constant 𝑘, the horizontal plane 𝑧 = 𝑘 intersects the surface 𝑧 = 𝑓(𝑥, 𝑦)
along a curve, called a contour curve of height 𝒌, whose equation is 𝑓(𝑥, 𝑦) = 𝑘.
The projection of the contour curve onto the 𝑥 − 𝑦 plane is a level curve of 𝑓. Hence, a level curve
consists of the set of points (𝑥, 𝑦) for which 𝑓(𝑥, 𝑦) has a constant value.
An example:
Dr NG Wee Seng
Page 3
AY2018-19 Semester 1 MA1511 Engineering Calculus
Remark
Functions of three or more variables, 𝑧 = 𝑓(𝑥1 , 𝑥2 , … , 𝑥𝑛 ) can be defined on some domain, 𝐷 of
points (𝑥1 , 𝑥2 , … , 𝑥𝑛 ) in a 𝑛 − dimensinal space . 𝑥1 , 𝑥2 , … , 𝑥𝑛 are the independent variables and z
is the dependent variable. The graph of 𝑓 is the set of points (𝑥1 , 𝑥2 , … , 𝑥𝑛 , 𝑓(𝑥1 , 𝑥2 , … , 𝑥𝑛 ) ) in a
(n+1) - dimensional space.
Dr NG Wee Seng
Page 4
AY2018-19 Semester 1 MA1511 Engineering Calculus
The vertical plane whose equation is 𝑦 = 𝑏 intersects the surface 𝑧 = 𝑓(𝑥, 𝑦) along a curve on
the plane whose equation is 𝑧 = 𝑓(𝑥, 𝑏) (substitute 𝑦 = 𝑏 in = 𝑓(𝑥, 𝑦) ).
The gradient of the tangent line to this curve at the point 𝑃(𝑎, 𝑏, 𝑓(𝑎, 𝑏)) is called the partial
derivative of 𝒇 at (𝑎, 𝑏) with respect to 𝒙.
We denote this quantity by 𝑓𝑥 (𝑎, 𝑏). The subscript 𝑥 indicates that 𝑥 is the variable.
Mathematically,
𝑓(𝑎 + ℎ, 𝑏) − 𝑓(𝑎, 𝑏)
𝑓𝑥 (𝑎, 𝑏) = lim
ℎ→0 ℎ
𝑓𝑥 (𝑎, 𝑏) gives the rate of change of 𝑓 at the point P (𝑎, 𝑏, 𝑓(𝑎, 𝑏)) in the direction of the 𝑥-axis.
Dr NG Wee Seng
Page 5
AY2018-19 Semester 1 MA1511 Engineering Calculus
The vertical plane whose equation is 𝑥 = 𝑎 intersects the surface 𝑧 = 𝑓(𝑥, 𝑦) along a curve on
the plane whose equation is 𝑧 = 𝑓(𝑎, 𝑦) (substitute 𝑥 = 𝑎 in = 𝑓(𝑥, 𝑦) ).
The gradient of the tangent line to this curve at the point 𝑃(𝑎, 𝑏, 𝑓(𝑎, 𝑏)) is called the partial
derivative of 𝒇 at (𝑎, 𝑏) with respect to 𝒚.
We denote this quantity by 𝑓𝑦 (𝑎, 𝑏). The subscript 𝑦 indicates that 𝑦 is the variable.
Mathematically,
𝑓(𝑎, 𝑏 + ℎ) − 𝑓(𝑎, 𝑏)
𝑓𝑦 (𝑎, 𝑏) = lim
ℎ→0 ℎ
𝑓𝑦 (𝑎, 𝑏) gives the rate of change of 𝑓 at the point where 𝑃 (𝑎, 𝑏, 𝑓(𝑎, 𝑏)) in the direction of the 𝑦-axis
Notations
𝜕𝑓
The partial derivatives , 𝑓𝑥 (𝑎, 𝑏) and 𝑓𝑦 (𝑎, 𝑏) are also commonly denoted by 𝜕𝑥 (𝑎, 𝑏) and
𝜕𝑓
(𝑎, 𝑏) respectively. As a function of 𝑥 and 𝑦, with no reference to any specific point, the partial
𝜕𝑦
𝜕𝑓 𝜕𝑓
derivatives can be written as 𝑓𝑥 , , 𝑓𝑦 , .
𝜕𝑥 𝜕𝑦
Dr NG Wee Seng
Page 6
AY2018-19 Semester 1 MA1511 Engineering Calculus
𝜕𝑓
• can be found by differentiating 𝑓(𝑥, 𝑦) with respect to 𝑦, treating 𝑥 as a constant
𝜕𝑦
𝜕𝑓
For example, if 𝑓(𝑥, 𝑦) = 2𝑥 2 + 3𝑦 2 , then = 6𝑦 since the partial derivative of 3𝑦 2 with
𝜕𝑦
𝜕𝑓
respect to 𝑦 is 6𝑦 while the partial derivative of 2𝑥 2 with respect to y is zero. Similarly, = 4𝑥.
𝜕𝑥
Remarks
Partial derivatives for functions of three or more variables can be defined and determined in a similar
fashion.
For example, if 𝑓(𝑥, 𝑦, 𝑧) = 2𝑥 + 𝑦 2 𝑧, then 𝑓𝑥 = 2, 𝑓𝑦 = 2𝑦𝑧 and 𝑓𝑧 = 𝑦 2 .
Dr NG Wee Seng
Page 7
AY2018-19 Semester 1 MA1511 Engineering Calculus
Differentiation of these functions produces the so-called second order partial derivatives of f
Notations we use are:
𝜕2 𝑓 𝜕 𝜕𝑓 𝜕𝑓
or 𝑓𝑥𝑥 denotes ( ) the partial derivative of w.r.t. 𝑥
𝜕𝑥 2 𝜕𝑥 𝜕𝑥 𝜕𝑥
𝜕2 𝑓 𝜕 𝜕𝑓 𝜕𝑓
or 𝑓𝑥𝑦 denotes ( ) the partial derivative of w.r.t. 𝑦
𝜕𝑦𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑥
𝜕2 𝑓 𝜕 𝜕𝑓 𝜕𝑓
or 𝑓𝑦𝑥 denotes ( ) the partial derivative of w.r.t. 𝑥
𝜕𝑥𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑦
𝜕2 𝑓 𝜕 𝜕𝑓 𝜕𝑓
or 𝑓𝑦𝑦 denotes ( ) the partial derivative of w.r.t. 𝑦
𝜕𝑦 2 𝜕𝑦 𝜕𝑦 𝜕𝑦
Dr NG Wee Seng
Page 8
AY2018-19 Semester 1 MA1511 Engineering Calculus
Remarks
1. Repeated differentiation of second order partial derivatives will produce partial derivatives of
order 3 and above. For example,
𝜕3 𝑓 𝜕 𝜕2 𝑓 𝜕2 𝑓
or 𝑓𝑥𝑦𝑦 denotes ( ) , the partial derivative of w.r.t. 𝑦
𝜕𝑦𝜕𝑦𝜕𝑥 𝜕𝑦 𝜕𝑦𝜕𝑥 𝜕𝑦𝜕𝑥
2. Higher order derivatives for functions of three or more variables can be defined in similar fashion.
For example, 𝑓(𝑥, 𝑦, 𝑧) = 2𝑥 + 𝑦 2 𝑧, then 𝑓𝑥 = 2, 𝑓𝑦 = 2𝑦𝑧 and 𝑓𝑧 = 𝑦 2 , so
𝜕 2𝑓 𝜕 𝜕𝑓 𝜕
or 𝑓𝑦𝑧 = ( )= ( 2𝑧𝑦) = 2𝑦
𝜕𝑧𝜕𝑦 𝜕𝑧 𝜕𝑦 𝜕𝑧
and so on.
Dr NG Wee Seng
Page 9
AY2018-19 Semester 1 MA1511 Engineering Calculus
Result 1.4A
A vector equation of the tangent plane at 𝑃 is
𝑓𝑥 (𝑎, 𝑏) 𝑓𝑥 (𝑎, 𝑏) 𝑎
𝐫 (𝑓𝑦 (𝑎, 𝑏)) = (𝑓𝑦 (𝑎, 𝑏)) ( 𝑏 )
−1 −1 𝑓(𝑎, 𝑏)
𝑥
where 𝐫 = (𝑦).
𝑧
Equivalently, a Cartesian equation of the plane is
𝑧 = 𝑓𝑥 (𝑎, 𝑏) ∙ (𝑥 − 𝑎) + 𝑓𝑦 (𝑎, 𝑏) ∙ (𝑦 − 𝑏) + 𝑓(𝑎, 𝑏)
Result 1.4B
The equation of the normal line at 𝑃 (which is the line passing through P and perpendicular to the
tangent plane ) is
𝑥 𝑎 𝑓𝑥 (𝑎, 𝑏)
(𝑦) = ( 𝑏 ) + (𝑓𝑦 (𝑎, 𝑏)) 𝑡, 𝑡 ∈ 𝐑
𝑧 𝑓(𝑎, 𝑏) −1
Dr NG Wee Seng
Page 10
AY2018-19 Semester 1 MA1511 Engineering Calculus
𝑥 2 −12
Dr NG Wee Seng
Page 11
AY2018-19 Semester 1 MA1511 Engineering Calculus
One way to remember result 1.5A is to draw the following tree diagram :
Dr NG Wee Seng
Page 12
AY2018-19 Semester 1 MA1511 Engineering Calculus
𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑦
= × + ×
𝜕𝑡 𝜕𝑥 𝜕𝑡 𝜕𝑦 𝜕𝑡
One way to remember result 1.5B is to draw the following tree diagrams:
𝝏𝒛 𝝏𝒛
𝒅𝒔 𝒅𝒕
Given that 𝑧 = 3𝑥 2 + 𝑦 2 , 𝑥 = 𝑡 3 −𝑒 𝑡 and 𝑦 = 𝑒 𝑡 + 𝑡 , use the chain rule to find the value of
𝑑𝑧
when 𝑡 = 0
𝑑𝑡
Answer: 10
Dr NG Wee Seng
Page 13
AY2018-19 Semester 1 MA1511 Engineering Calculus
2 2
Answer: 8𝑠𝑡 − 6𝑡 ; 4𝑠 − 12𝑠𝑡
Dr NG Wee Seng
Page 14
AY2018-19 Semester 1 MA1511 Engineering Calculus
Given a surface 𝑧 = 𝑓(𝑥, 𝑦), a point 𝑃(𝑎, 𝑏, 𝑓(𝑎, 𝑏)) on the surface and a direction unit vector u =
𝑢1 i + 𝑢2 j , the vertical plane containing P and parallel to u intersects the surface along a curve . The
gradient of the tangent to this curve at 𝑃 is called the directional derivative of 𝑓(𝑥, 𝑦) at (𝑎, 𝑏) in
the direction of u. We denote this derivative by 𝐷𝐮 𝑓(𝑎, 𝑏). We can interpret 𝐷𝐮 𝑓(𝑎, 𝑏) as the rate of
change of the function 𝑓(𝑥, 𝑦) in the direction of the vector u.
tangent at P
f (a hu1 , b hu 2 ) f (a, b)
𝐷𝐮 𝑓(𝑎, 𝑏) = lim
h0 h
The following result provides us with an efficient way to compute directional derivatives.
Result 1.6A
For any unit vector u = 𝑢1 i + 𝑢2 j ,
𝑓𝑥 (𝑎, 𝑏) 𝑢1
𝐷𝐮 𝑓(𝑎, 𝑏) = 𝑢1 𝑓𝑥 (𝑎, 𝑏) + 𝑢2 𝑓𝑦 (𝑎, 𝑏) = ( ) (𝑢 )
𝑓𝑦 (𝑎, 𝑏) 2
Dr NG Wee Seng
Page 15
AY2018-19 Semester 1 MA1511 Engineering Calculus
Remarks
1. If the given direction vector 𝑢1 i + 𝑢2 j is not a unit vector, we can normalise the vector to
get a unit vector:
1
u= (𝑢1 i + 𝑢2 j )
√𝑢2 1 +𝑢 2
2
2. The gradient vector of 𝑓 , denoted by ∇𝑓, is the vector
𝑓𝑥
𝑓𝑥 𝐢 + 𝑓𝑦 𝐣 (= ( ))
𝑓𝑦
𝐷𝐮 𝑓(𝑎, 𝑏) = ∇𝑓(𝑎, 𝑏) 𝐮
3. The first order partial derivatives of 𝑓 are special cases of directional derivatives. Specifically,
4. Directional derivatives for functions of three or more variables can be defined in a similar
fashion.
𝑓𝑥 (𝑎, 𝑏, 𝑐) 𝑢1
𝐷𝐮 𝑓(𝑎, 𝑏, 𝑐) = 𝑢1 𝑓𝑥 (𝑎, 𝑏, 𝑐) + 𝑢2 𝑓𝑦 (𝑎, 𝑏, 𝑐) + 𝑢3 𝑓𝑧 (𝑎, 𝑏, 𝑐) = (𝑓𝑦 (𝑎, 𝑏, 𝑐)) ( 2 )
𝑢
𝑓𝑧 (𝑎, 𝑏, 𝑐) 𝑢3
Dr NG Wee Seng
Page 16
AY2018-19 Semester 1 MA1511 Engineering Calculus
−4 4
13
Answers (i) − (ii) (√17 √17
1 ) , ( −1 )
5
√17 √17
Dr NG Wee Seng
Page 17
AY2018-19 Semester 1 MA1511 Engineering Calculus
Local Maximum
𝑓 has a local maximum at (𝑎, 𝑏) if 𝑓(𝑎, 𝑏) ≥ 𝑓(𝑥, 𝑦) for all (𝑥, 𝑦) in some open subset of 𝐷
containing (𝑎, 𝑏)
Local Minimum
𝑓 has a local minimum at (𝑎, 𝑏) if 𝑓(𝑎, 𝑏) ≤ 𝑓(𝑥, 𝑦) for all (𝑥, 𝑦) in some open subset of 𝐷
containing (𝑎, 𝑏)
We recall from single−variable calculus that if a differentiable function 𝑓(𝑥) in one variable has a
local maximum/minimum at 𝑥 = 𝑐, then 𝑓 ′ (𝑐) = 0. An analogous result holds for functions in two
variables.
Result 1.7A
If 𝑓 has a local maximum/minimum at an interior point (𝑎, 𝑏) of its domain, then
𝑓𝑥 (𝑎, 𝑏) = 0 and 𝑓𝑦 (𝑎, 𝑏) = 0
Critical Point
Dr NG Wee Seng
Page 18
AY2018-19 Semester 1 MA1511 Engineering Calculus
Result 1.7A says that a local maximum or a local minimum occurs only at a critical point. Hence, to
locate local maximum or local minimum of 𝑓(𝑥, 𝑦), we will solve the simultaneous equations
𝑓𝑥 = 0 and 𝑓𝑦 = 0
for all possible critical points. The nature (local maximum or local minimum, or neither) can be
determined using the following derivative test
Remark
The above test is inconclusive if 𝐷 = 0
Dr NG Wee Seng
Page 19
AY2018-19 Semester 1 MA1511 Engineering Calculus
Dr NG Wee Seng
Page 20
AY2018-19 Semester 1 MA1511 Engineering Calculus
Dr NG Wee Seng
Page 21
AY2018-19 Semester 1 MA1511 Engineering Calculus
Given a function 𝑓(𝑥, 𝑦), we seek the extreme values of 𝑓 when (𝑥, 𝑦) is subject to one constraint
given by an equation of the form
𝑔(𝑥, 𝑦) = 0
Geometrically, we find the maximum/minimum values of 𝑓(𝑥, 𝑦) for points (𝑥, 𝑦) lying on the curve
whose equation is 𝑔(𝑥, 𝑦) = 0.
To find largest/smallest
value of f along this curve
Result 1.8A
Dr NG Wee Seng
Page 22
AY2018-19 Semester 1 MA1511 Engineering Calculus
Remarks
Geometrically, the equations
𝑓𝑥 = λ𝑔𝑥 , 𝑓𝑦 = λ𝑔𝑦 ,
indicate that the maximum or minimum occurs at a point where the gradient vector of 𝑓 is parallel to
the gradient vector of 𝑔, as depicted in the diagram below.
Dr NG Wee Seng
Page 23
AY2018-19 Semester 1 MA1511 Engineering Calculus
Dr NG Wee Seng
Page 24
AY2018-19 Semester 1 MA1511 Engineering Calculus
𝑔(𝑥1 , 𝑥2 , … , 𝑥𝑛 ) = 0
for some constant λ , called the Lagrange multiplier
End of Chapter 1
“Mathematics is the art of giving the same name to different things.” – Henri Poincaré
Dr NG Wee Seng