Vector Calculus

Lecture IV

Albena Veltcheva
Waseda University
Summer Quarter 2025
1
Gradient, Divergence, Curl, and
the Del Operator

Gradient

2
 Gradient
The gradient is defined for a scalar function Φ(x,y,z) as:

𝜕 𝜕 𝜕 𝜕Φ 𝜕Φ 𝜕Φ
𝛻Φ = 𝑖Ԧ + 𝑗Ԧ + 𝑘 Φ = 𝑖Ԧ + 𝑗Ԧ +𝑘
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧

𝛻𝛷, grad 𝛷– vector-valued function

The gradient itself is a vector function pointing to the

direction with the largest increase per distance
moved for the scalar function Φ
For scalar fields Φ, Ψ: 𝛻 Φ + Ψ = 𝛻Φ + 𝛻Ψ 3
 Applications of Gradient
➢ Directional Derivatives
➢ Critical points
➢ Spherical gradient, Gradient field and
potential
➢ Lagrange Multiplier
➢ Normal Vector

4
 Lagrange Multipliers
Constrained Extrema

Abstract Statement of Problem

● Find extrema of f (x1,…, xn) subject to a
constraint g(x1,…, xn) = c.

5
Copyright 2012 Pearson Education, Inc.
 Lagrange Multiplier Method
➢ We are looking for extrema of 𝑓 𝑥, 𝑦 while
“moving” on g 𝑥, 𝑦

➢ A level curve of 𝑓 𝑥, 𝑦 is one that is

perpendicular to 𝛻𝑓 and therefore at an
extremum, 𝛻𝑓must be parallel to 𝛻𝑔

➢ → 𝛻𝑓 at an extremum is proportional to 𝛻𝑔
𝛻𝑓 parallel to 𝛻𝑔:
𝛻𝑓 = 𝜆𝛻𝑔
 - Lagrange Multiplier 6
 Lagrange Multiplier Theorem

● Let f : X → R and g: X → R be functions, where

X  Rn is open.
● Let S = {x  Rn | g(x) = c}.
● If f |S has an extremum at a point a  S where
g(a) ≠ 0, then there must be a real number 
such that
f (a) = g(a).
 - Lagrange Multiplier
7
Copyright 2012 Pearson Education, Inc.
Finding Constrained Critical Points

To find the constrained critical points of f

subject to g(x) = c, solve

f (x) = g(x)
g(x) = c
really n
for x, . equations

8
Copyright 2012 Pearson Education, Inc.
 Exercise
Find extrema of a function 𝑓 𝑥, 𝑦 = 𝑥 2 + 2𝑦 2 under the
constraint 𝑔 𝑥, 𝑦 = 2𝑥 + 𝑦 = 9

𝟐𝒙 + 𝒚 = 𝟗
Find the position and value of this minimum

9
 Solution

𝑓 𝑥, 𝑦 = 𝑥 2 + 2𝑦 2
𝑔 𝑥, 𝑦 = 2𝑥 + 𝑦 = 9
𝛻𝑓 = λ𝛻𝑔
2𝑥 = 2λ
4𝑦 = λ
2𝑥 + 𝑦 = 9

𝑥 = 4𝑦 , 8𝑦 + 𝑦 = 9, 𝑦 = 1 , 𝑥 = 4
𝟐𝒙 + 𝒚 = 𝟗

Only one solution, the minimum

𝑓 4,1 = 16 + 2 = 18 10
 Exercise
➢ find f(x,y,z) and g(x,y,z)
➢ create system of equations
to solve
y=? 54
z=? (3D → 4 equations)
x=? ➢ find solutions to equations
➢ discard those solutions for
Calculate the edge lengths which the box has zero or
x,y,z of a box with negative volume
maximum volume, given a
➢ λ = 0 solutions discard the
total surface area of 54
parallelism of the gradients
[area unit] → methods initial condition
no longer fulfilled
11
 Solution

Volume:
𝑓 𝑥, 𝑦, 𝑧 = 𝑥𝑦𝑧
Area / 2 = 27:
𝑔 𝑥, 𝑦, 𝑧 = 2𝑥𝑦 + 2𝑥𝑧 + 2𝑦𝑧 = 54 𝑔 𝑥, 𝑦, 𝑧 = 𝑥𝑦 + 𝑥𝑧 + 𝑦𝑧 = 27

𝑦𝑧 𝑦+𝑧
𝛻𝑓 = 𝑥𝑧 , 𝛻𝑔 = 𝑥 + 𝑧
𝑥𝑦 𝑥+𝑦
𝑦𝑧 = λ 𝑦 + 𝑧
𝑥𝑧 = λ 𝑥 + 𝑧 System of
𝑥𝑦 = λ 𝑥 + 𝑦 Equations
𝑥𝑦 + 𝑥𝑧 + 𝑦𝑧 = 27

12
 Solution

λ𝑥 𝑦 + 𝑧 = λ𝑦 𝑥 + 𝑧
𝑦𝑧 = λ 𝑦 + 𝑧
λ 𝑥𝑦 + 𝑥𝑧 − 𝑥𝑦 + 𝑦𝑧 = 0
𝑥𝑧 = λ 𝑥 + 𝑧
λ 𝑥𝑧 − 𝑦𝑧 = 0
𝑥𝑦 = λ 𝑥 + 𝑦
𝑥𝑦 + 𝑥𝑧 + 𝑦𝑧 = 27 λ = 0 , 𝑥𝑧 = 𝑦𝑧 , 𝑧 = 0 , 𝑥 = 𝑦

multiply first 3 equations λ𝑦 𝑥 + 𝑧 = λ𝑧 𝑥 + 𝑦

with x / y / z so that the λ 𝑥𝑦 + 𝑦𝑧 − 𝑥𝑧 + 𝑦𝑧 = 0
left sides become equal: λ 𝑥𝑦 − 𝑥𝑧 = 0
𝑥𝑦𝑧 = λx 𝑦 + 𝑧
λ = 0 , 𝑥𝑦 = 𝑥𝑧 , 𝑥 = 0 , 𝑦 = 𝑧
𝑥𝑦𝑧 = λ𝑦 𝑥 + 𝑧
𝑥𝑦z = λz 𝑥 + 𝑦 𝑥=𝑦=𝑧
𝑥 2 + 𝑥 2 + 𝑥 2 = 3𝑥 2 = 27
λ= 0 : gradients being parallel not guaranteed 27
x, y, z = 0 : box volume = 0 𝑥=𝑦=𝑧=± =3
x, y ,z can not be negative 3 13
 Exercise

Find extrema of a function 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦 under the

constraint 𝑔 𝑥, 𝑦 = 𝑥 2 + 𝑦 2 = 1

14
 Solution
Find extrema of a function 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦 under the
constraint 𝑔 𝑥, 𝑦 = 𝑥 2 + 𝑦 2 = 1
𝛻𝑓 = 𝜆𝛻𝑔
2𝑥 2𝑥
𝛻𝑓 = , 𝛻𝑔 =
1 2𝑦
2𝑥 = λ2𝑥
1 = λ2𝑦
𝑥2 + 𝑦2 = 1

15
 Solution
2𝑥 = λ2𝑥 2𝑥 = λ2𝑥
1 = λ2𝑦
λ=1 OR 𝑥=0
𝑥2 + 𝑦2 = 1
1 𝑦2 = 1
𝑦=
2
1 𝑦 = 1 𝑦 = −1
2
𝑥 + =1
4
𝑥=0 𝑥=0
𝑥 = 0𝑦 =1
3 3 𝑥 = 0 𝑦 = −1
𝑥=
4 𝑥=− 4 3 1
𝑥= 𝑦=
4 2 16

3 1
𝑥=− 𝑦=
4 2
 Solution
The obtained points are only critical points:
→ may be minimum, maximum or saddle point
𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦

𝑥 = 0, 𝑦 = 1 → 𝑓 0,1 = 1 Local Minimum

𝑥 = 0, 𝑦 = −1 → 𝑓 0, −1 = −1 Global Minimum

3 1 3 1 5
𝑥= ,𝑦 = → 𝑓 , = Global Maximum
4 2 4 2 4

3 1 3 1 5
𝑥 = − ,𝑦 = → 𝑓 − 4,2 = 4 Global Maximum
4 2

Since the constraint defines a closed, bounded curve, by the

extreme value theorem f must attain both a global maximum
17
and a global minimum on this curve.
 Global Extrema on Compact Regions
Reminder

How to identify global extrema?

- There is no general, effective method for looking at an
arbitrary function and determining whether and where it
reaches an absolute maximum or minimum value.
- We can say something about global extrema for functions
defined on a certain type of domain – compact regions

18
 Extrema on Compact
Regions(reminder)
● Definition A region (subset) D in Rn is compact if it is
closed (i.e., contains all its boundary points) and
bounded (i.e., can be enclosed in a suitable large n-
dimensional ball).
● ➔ compact sets contain their boundaries (a consequence
of being closed) and have only finite extent (a
consequence of being bounded).

19
 Extreme Value Theorem (reminder)
The notion of compactness is of value because of the socalled
Extreme Value theorem

Extreme Value Theorem

If D is a compact region in Rn and f : D → R is
continuous, then f must have a (global) maximum and
minimum values on D.
That is, there must exist points 𝒂𝑚𝑎𝑥 and 𝒂𝑚𝑖𝑛 in D
such that, for all x ∈ D,
𝑓(𝒂𝑚𝑖𝑛 ) ≤ 𝑓(𝒙) ≤ 𝑓(𝒂𝑚𝑎𝑥 )
20
 Finding Global Extrema on Compact
Regions (reminder)
Thus, to find the global maximum and minimum of a
differentiable function f on a compact region D:
1. Find any “ordinary” critical points of f occurring on D.
2. Find any “restricted” critical points of f occurring on D
(the boundary of D).
3. Compare values of f at all points found in #1 and #2.

If a function is defined over a noncompact region, there is no general result

like the extreme value theorem to guarantee existence of any global extrema.
However, ad hoc arguments frequently can be used to identify global extrema.
21
 Exercise
Find the extrema of 𝑓 𝑥, 𝑦 = 2𝑥 + 𝑦 subject to the conditions
that 𝑥 + 𝑦 = 3

22
 Solution
Find the extrema of 𝑓 𝑥, 𝑦 = 2𝑥 + 𝑦 subject to the conditions
that 𝑥 + 𝑦 = 3
Let 𝑔 𝑥, 𝑦 = 𝑥 + 𝑦 = 3
∇𝑓 𝑥, 𝑦 = 𝜆∇𝑔 𝑥, 𝑦
2 = 𝜆/(2 𝑥)
൞ 1 = 𝜆/(2 𝑦)
𝑥+ 𝑦=3
𝜆
2= ⇒𝜆=4 𝑥=2 𝑦
2 𝑥
𝑦=2 𝑥
𝑥+ 𝑦=3⇒3 𝑥=3⇒𝑥=1
Then 𝑦 = 2 𝑥 ⇒ 𝑦 = 4
One critical point 1,4 23
 Solution
Find the extrema of 𝑓 𝑥, 𝑦 = 2𝑥 + 𝑦 subject to the conditions
that 𝑥 + 𝑦 = 3 One critical point 1,4
Since the constraint defines a closed, bounded
curve segment, by the extreme value theorem
f must attain both a global maximum and a
global minimum on this segment.
By the Lagrange multiplier method only a
single critical point 1,4 . However the points
(9, 0) and (0, 9) satisfy the constraint
𝑥 + 𝑦 = 3. But they are both points where
∇𝑔 is undefined
f=? at (1,4), (9,0) and (0,9)
f (1, 4) = 6, while f (9, 0) = 18 and f (0, 9) = 9.
Comparing the value of 𝑓 𝑥, 𝑦 at these three points we conclude24
that the minimum of f occurs at (1, 4) and the maximum at (9, 0).
 Solution
Find the extrema of 𝑓 𝑥, 𝑦 = 2𝑥 + 𝑦 subject to the conditions that 𝑥 + 𝑦 = 3
The minimum of f occurs at (1, 4) and the
maximum at (9, 0).
Geometry: The level sets of the function f =
the family of parallel lines 2𝑥 + 𝑦 = 𝑘. The
height 𝑘 is the y-intercept of the corresponding
line in the family. → looking to find the largest
and smallest y-intercepts of any line in the
family that meets the curve 𝑥 + 𝑦 = 3
These extreme values of 𝑘 occur either when one of the lines is
tangent to the constraint curve or at an endpoint of the curve.
This example – illustration of the importance of locating all the
points where extrema may occur by considering places where
∇𝑓 or ∇𝑔 is undefined (or where ∇𝑔 = 𝟎) as well as the critical
points determined by Lagrange multiplier method. 25
 Lagrange Multipliers
More than one constrain
Generalization
Find extrema of a function f subject to several constraints.

26
 More Than One Constraint

● Find extrema of f (x1,…, xn) subject to

g1(x1,…, xn) = c1
g2(x1,…, xn) = c2

gk(x1,…, xn) = ck

where k < n.
27
 Theorem
● Let f : X → R and g1,…, gk: X → R be functions of class
C1, where X  Rn is open.
● Let S ={x  Rn | g1(x) = c1,…, gk(x) = ck} are
constrains.
● If f |S has an extremum at a point a  S where
g1(a),…,gk(a) are linearly independent*, then there
exist scalars 1,…, k such that
f (a) = 1g1(a) + 2g2(a) + … + kgk(a).
k vectors v1,…,vk are linearly independent if the only way to satisfy
equation a1v1 +… + akvk = 0 for scalars a1, … ak is if
a1 = a2 … = ak = 0. 28
 Finding Constrained
Critical Points

● To find constrained critical points of f subject to

g1(x) = c1
solve
g2(x) = c2,
f (x) = 1g1(x) +…+ kgk(x)


gk(x) = ck g1(x) = c1


gk(x) = ck
for x, 1,…, k.
29
 Exercise
Use Lagrange multipliers to identify the critical points of
𝑓 𝑥, 𝑦, 𝑧 = 2𝑥 + 𝑦 2 − 𝑧 2 , subject to the constrains 𝑥 − 2𝑦 = 0,
𝑥+𝑧 =0

30
 Solution
Use Lagrange multipliers to identify the critical points of
𝑓 𝑥, 𝑦, 𝑧 = 2𝑥 + 𝑦 2 − 𝑧 2 , subject to the constrains 𝑥 − 2𝑦 = 0,
𝑥+𝑧 =0
𝑔1 (𝑥, 𝑦, 𝑧) = 𝑥 − 2𝑦 = 0
ቊ
𝑔2 (𝑥, 𝑦, 𝑧) = 𝑥 + 𝑧 = 0
By the Theorem, any constrained critical points (𝑥0 , 𝑦0 , 𝑧0 ) must
satisfy ∇𝑓 𝑥0 , 𝑦0 , 𝑧0 = 𝜆1 ∇𝑔1 𝑥0 , 𝑦0 , 𝑧0 + 𝜆2 ∇𝑔2 𝑥0 , 𝑦0 , 𝑧0
and the two constrained →system
2 = 𝜆1 + 𝜆2
2𝑦 = −2𝜆1
−2𝑧 = 𝜆2
𝑔1 (𝑥, 𝑦, 𝑧) = 𝑥 − 2𝑦 = 0
𝑔2 (𝑥, 𝑦, 𝑧) = 𝑥 + 𝑧 = 0

31
 Solution
2 = 𝜆1 + 𝜆2
2𝑦 = −2𝜆1
−2𝑧 = 𝜆2
𝑔1 (𝑥, 𝑦, 𝑧) = 𝑥 − 2𝑦 = 0
𝑔2 (𝑥, 𝑦, 𝑧) = 𝑥 + 𝑧 = 0
Subtract from 3rd (x2) eq 2nd one and compare with 1st
2 = 𝜆1 + 𝜆2 = −𝑦 − 2𝑧
Substitute 𝑧 and 𝑦 from 4th and 5th
𝑥 4
2 = −𝑦 − 2𝑧 = − + 2𝑥 ⇒ 𝑥 =
2 3
4 2 4
The critical point is , , −
3 3 3

32
 Exercise
Use Lagrange multipliers to identify the critical points of
𝑓 𝑥, 𝑦, 𝑧 = 2𝑥 + 𝑦 2 + 2𝑧, subject to the constrains 𝑥 2 − 𝑦 2 = 1,
𝑥+𝑦+𝑧 =2

33
 Solution
Use Lagrange multipliers to identify the critical points of
𝑓 𝑥, 𝑦, 𝑧 = 2𝑥 + 𝑦 2 + 2𝑧, subject to the constrains 𝑥 2 − 𝑦 2 = 1,
𝑥+𝑦+𝑧 =2
𝑔1 (𝑥, 𝑦, 𝑧) = 𝑥 2 − 𝑦 2 = 1
ቊ
𝑔2 𝑥, 𝑦, 𝑧 = 𝑥 + 𝑦 + 𝑧 = 2
By the Theorem, any constrained critical points (𝑥0 , 𝑦0 , 𝑧0 ) must
satisfy ∇𝑓 𝑥0 , 𝑦0 , 𝑧0 = 𝜆1 ∇𝑔1 𝑥0 , 𝑦0 , 𝑧0 + 𝜆2 ∇𝑔2 𝑥0 , 𝑦0 , 𝑧0
and the two constrained →system
2 = 2𝑥𝜆1 + 𝜆2
2𝑦 = −2𝑦𝜆1 + 𝜆2
2 = 𝜆2
𝑔1 (𝑥, 𝑦, 𝑧) = 𝑥 2 − 𝑦 2 = 1
𝑔2 (𝑥, 𝑦, 𝑧) = 𝑥 + 𝑦 + 𝑧 = 2

34
 Solution
2 = 2𝑥𝜆1 + 𝜆2
2𝑦 = −2𝑦𝜆1 + 𝜆2
2 = 𝜆2
𝑔1 (𝑥, 𝑦, 𝑧) = 𝑥 2 − 𝑦 2 = 1
𝑔2 (𝑥, 𝑦, 𝑧) = 𝑥 + 𝑦 + 𝑧 = 2
2 = 𝜆2 ⇒ 2𝑥𝜆1 = 0 ⇒ 𝜆1 = 0 or 𝑥 = 0
If 𝑥 = 0 ⇒ 𝑥 2 − 𝑦 2 = 1 ⇒ −𝑦 2 = 1 – no solution
If 𝜆1 = 0 ⇒ 2𝑦 = 2 ⇒ 𝑦 = 1
From 4th eq 𝑥 2 = 2 ⇒ 𝑥 = ± 2
From 5th eq. 𝑧 = 1 ∓ 2
The critical points are 2, 1,1 − 2 and − 2, 1,1 + 2

35
 Hessian Criterion for
Constrained Extrema
Optional
Although sometimes the determination of the nature of a critical
point (constrained or unconstrained) can be done from
considerations it is better to have a more automatic way to find
the type of a constrained critical point.
→ Hessian criterion for constrained critical points
Like the one in the unconstrained case, Hessian criterion only
determines the local nature of a critical point and no information
about global constrained extrema

36
 The Hessian Criterion: Set-up
Find extrema of a function f : X ⊆ Rn → R subject to the k
constraints
g1(x1,…, xn) = c1
g2(x1,…, xn) = c2


gk(x1,…, xn) = ck
Assume that g1(x),…gk(x) are linearly independent at the
constrained critical point a. By the theorem, any constrained
extremum a must satisfy
f (x) = 1g1(x) +…+ kgk(x)
for some scalars 1, . . . , k . Then we can consider a constrained
critical point to be a pair of vectors (; a) = (1,…, k; a1,…, an)
37
 The Hessian Criterion: Set-up
● Consider a constrained critical point of f subject to
● g1 = c1,…, gk = ck to be a pair of vectors
● (; a) = (1,…, k; a1,…, an)
● In fact, (; a) is an unconstrained critical point of the
function 𝐿 defined by
● 𝐿 𝑙1 , 𝑙2 , … , 𝑙𝑘 ; 𝑥1 , 𝑥2 , … , 𝑥𝑛 = 𝑓 𝑥1 , 𝑥2 , … , 𝑥𝑛
𝑘
σ
− 𝑖=1 𝑙𝑖 𝑔𝑖 𝑥1 , 𝑥2 , … , 𝑥𝑛 − 𝑐𝑖

Lagrangian function L
 The Hessian Criterion: Set-up
The Hessian criterion comes from considering the
Hessian of L at the critical point (λ; a).
𝐿 𝑙1 , 𝑙2 , … , 𝑙𝑘 ; 𝑥1 , 𝑥2 , … , 𝑥𝑘
𝑘

= 𝑓 𝑥1 , 𝑥2 , … , 𝑥𝑘 − ෍ 𝑙𝑖 𝑔𝑖 𝑥1 , 𝑥2 , … , 𝑥𝑘 − 𝑐𝑖
𝑖=1

◼ Let hij denote 2L/xj xi evaluated at (; a):

2 2 2
 f  g1  gk
hij = (a) − 1 (a) −  − k (a).
x j xi x j xi x j xi
 The Hessian Criterion: Set-up
From linear algebra: Since g1(x),…gk(x) are linearly independent
at the constrained critical point a. Then derivative matrix of
g = (g1, . . . , gk) at a,
𝜕𝑔1 𝜕𝑔1
𝒂 … 𝒂
𝜕𝑥1 𝜕𝑥𝑛
𝑫𝒈 𝒂 = ⋮ ⋱ ⋮
𝜕𝑔𝑘 𝜕𝑔𝑘
𝒂 𝒂
𝜕𝑥1 𝜕𝑥𝑛
has a k × k submatrix (obtained by deleting n − k columns of Dg(a)) with
nonzero determinant. By relabeling the variables, we will assume
𝜕𝑔1 𝜕𝑔1
𝒂 … 𝒂
𝜕𝑥1 𝜕𝑥𝑘
𝑑𝑒𝑡 ⋮ ⋱ ⋮ ≠0
𝜕𝑔𝑘 𝜕𝑔𝑘
𝒂 𝒂
𝜕𝑥1 𝜕𝑥𝑘
(i.e., that we may delete the last n − k columns).
 The Hessian Criterion: Set-up
Look at the Hessian HL(; a) of L at (; a):
k n

 0 g1 g1
 0 − (a)  − (a) 
 x1 xn 
k       
 g k g k 
 0  0 − (a)  − (a)
x1 xn
 g1 g k 
 − x1 (a)  −
x1
(a) h11  h1n 
       
n  g g k 
−
 x 1 (a)  − x (a) hn1  hnn 
 n n 
HL(; a) is (𝑛 + 𝑘) × (𝑛 + 𝑘) matrix
 The Hessian Criterion: Set-up

Assume the variables x1,…, xn are labeled so that

 g1 (a)  g1 (a) 

 x1 xk 
det       0.
 g k g k 
 x1 (a )  (a ) 
xk
 The Hessian Criterion: Set-up

◼ Set Hj = upper-leftmost j  j submatrix of HL(; a).

◼ Set hj = det Hj for j = 1, 2,…, k + n.
◼ Compute the sequence of numbers
(–1)kh2k+1, (–1)kh2k+2,…, (–1)khk+n. (1)
Note that, if k ≥ 1, the sequence in (1) is not the complete sequence
of principal minors of HL(λ, a).
 The Criterion
Assume that hk+n= det HL(; a) ≠ 0.
1. If all the numbers in (1) are positive, then f has a local minimum
at a subject to the constraints g1(x) = c1, g2(x) = c2,… gk(x) = ck
2. If the sequence (1) begins with a negative number, then
alternates in sign, then f has a local maximum at a subject to the
constraints g1(x) = c1, g2(x) = c2,… gk(x) = ck
3. If neither case 1 nor case 2 occurs, then f has a saddle point at a.
● If hk+n= det HL(; a) = 0, the critical point a is degenerate and
the test fails. (need to use another method to determine whether
or not it is extremum).
● In the case of no constraint equations gi(x) = ci (i.e., k = 0), the
criterion becomes the usual Hessian test for a function f of n
variables
 Exercise
An open rectangular box is to be manufactured having a (fixed)
volume of 4 m3. Use the Lagrange multipliers to find what
dimensions should the box have so as to minimize the amount of
material used to make it. Use the Hessian criterion to check that
the critical point is a local minimum.

45
 Solution
An open rectangular box is to be manufactured having a (fixed)
volume of 4 m3. Use the Lagrange multipliers to find what
dimensions should the box have so as to minimize the amount of
material used to make it.
➔to minimize the surface area
𝐴 𝑥, 𝑦, 𝑧 = 2𝑥𝑦 + 2𝑦𝑧 + 𝑥𝑧
subject to the constrain
𝑉 𝑥, 𝑦, 𝑧 = 𝑥𝑦𝑧 = 4
1. ∇𝐴 𝑥, 𝑦, 𝑧 = 𝜆∇𝑉 𝑥, 𝑦, 𝑧
2. System
2𝑦 + 𝑧 = 𝜆𝑦𝑧
2𝑥 + 2𝑧 = 𝜆𝑥𝑧
2𝑦 + 𝑥 = 𝜆𝑥𝑦
𝑥𝑦𝑧 = 4
46
 Solution
1. ∇𝐴 𝑥, 𝑦, 𝑧 = 𝜆∇𝑉 𝑥, 𝑦𝑧
2𝑦 + 𝑧 = 𝜆𝑦𝑧
2𝑥 + 2𝑧 = 𝜆𝑥𝑧
2. System 2𝑦 + 𝑥 = 𝜆𝑥𝑦
𝑥𝑦𝑧 = 4
2𝑦 + 𝑧 2𝑥 + 2𝑧 2𝑦 + 𝑥
𝜆= = =
𝑦𝑧 𝑥𝑧 𝑥𝑦
2 1 2 2 2 1
+ = + = +
𝑧 𝑦 𝑧 𝑥 𝑥 𝑦
2 1 2 2
+ = + ⇒ 𝑥 = 2𝑦
𝑧 𝑦 𝑧 𝑥
2 2 2 1
+ = + ⇒ 𝑧 = 2𝑦
𝑧 𝑥 𝑥 𝑦
𝑥𝑦𝑧 = 4 ⇒ 2𝑦 𝑦 2𝑦 = 4 ⇒ 𝑦 = 1
𝑥 = 2, 𝑦 = 1, 𝑧 = 2 ⇒ 𝜆 = 2 47
 Solution
One critical point 𝑥 = 2, 𝑦 = 1, 𝑧 = 2 ⇒ 𝜆 = 2
Lagrangian function L
𝐿 𝑙; 𝑥, 𝑦, 𝑧 = 𝐴 𝑥, 𝑦, 𝑧 − 𝑙 𝑉 𝑥, 𝑦, 𝑧 − 4
= 2𝑥𝑦 + 2𝑦𝑧 + 𝑥𝑧 − 𝑙(𝑥𝑦𝑧 − 4)
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝑙 2 𝜕𝑙𝜕𝑥 𝜕𝑙𝜕𝑦 𝜕𝑙𝜕𝑧
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝑥𝜕𝑙 𝜕𝑥 2 𝜕𝑥𝜕𝑦 𝜕𝑥𝜕𝑧
𝐻𝐿 𝑙; 𝑥, 𝑦, 𝑧 =
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝑦𝜕𝑙 𝜕𝑦𝜕𝑥 𝜕𝑦 2 𝜕𝑦𝜕𝑧
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝑧𝜕𝑙 𝜕𝑧𝜕𝑥 𝜕𝑧𝜕𝑦 𝜕𝑧 2
48
 Solution
One critical point 𝑥 = 2, 𝑦 = 1, 𝑧 = 2 ⇒ 𝜆 = 2
Lagrangian function L
𝐿 𝑙; 𝑥, 𝑦, 𝑧 = 𝐴 𝑥, 𝑦, 𝑧 − 𝑙 𝑉 𝑥, 𝑦, 𝑧 − 4
= 2𝑥𝑦 + 2𝑦𝑧 + 𝑥𝑧 − 𝑙(𝑥𝑦𝑧 − 4)
0 −𝑦𝑧 −𝑥𝑧 −𝑥𝑦
−𝑦𝑧 0 2 − 𝑙𝑧 1 − 𝑙𝑦
𝐻𝐿 𝑙; 𝑥, 𝑦, 𝑧 = −𝑥𝑧
2 − 𝑙𝑧 0 2 − 𝑙𝑥
−𝑥𝑦 1 − 𝑙𝑦 2 − 𝑙𝑥 0
0 −2 −4 −2
𝐻𝐿 2; 2,1,2 = −2 0 −2 −1
−4 −2 0 −2
−2 −1 −2 0
(–1)kh2k+1=?
0 −2 −4
(−1)1 det 𝐻2 1 +1 = −1 det −2 0 −2 = 32
−4 −2 0 49
 Solution
0 −2 −4 −2
𝐻𝐿 2; 2,1,2 = −2 0 −2 −1
−4 −2 0 −2
−2 −1 −2 0
(–1)kh2k+1=?
𝑘 = 1, 𝑛 = 3 ⇒
0 −2 −4
(−1)1 det 𝐻2 1 +1 = −1 det −2 0 −2 = 32
−4 −2 0
0 −2 −4 −2
(−1)1 det 𝐻1+3 = −1 det −2 0 −2 −1 = 48
−4 −2 0 −2
−2 −1 −2 0
Since the sequence of (–1)kh2k+1 are both positive, the
critical point (2, 1, 2) indeed minimizes the area of the box
subject to the constant volume=4.
50
 Exercise
a) Find the critical points of 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦, subject to the
constrain 𝑥 2 + 2𝑦 2 = 1
b) Use the Hessian criterion to determine the nature of the critical
point.

51
 Solution
a) Find the critical points of 𝑓 𝑥, 𝑦 = 𝑥 2 + 𝑦, subject to the constrain 𝑥 2
+ 2𝑦 2 = 1
Let 𝑔 𝑥, 𝑦 = 𝑥 2 + 2𝑦 2 = 1
∇𝑓 𝑥, 𝑦 = 𝜆∇𝑔 𝑥, 𝑦
2𝑥 = 2𝑥𝜆
ቐ 1 = 4𝑦𝜆
𝑥 2 + 2𝑦 2 = 1
From 1st eq. 𝑥 = 0 or 𝜆 = 1.
If 𝑥 = 0 ⇒ 𝑦 = ± 1/2
1
If 𝜆 = 1 ⇒ 𝑦 = ⇒ 𝑥 = ± 7/8
4
7 1
The critical points are ± , for 𝜆 = 1
8 4

1
and 0, ± for 𝜆 = 2/4
2 52
 Solution
b) Use the Hessian criterion to determine the nature of the critical
2𝑥 = 2𝑥𝜆
point. ∇𝑓 𝑥, 𝑦 = 𝜆∇𝑔 𝑥, 𝑦 ⇒ ቐ 1 = 4𝑦𝜆
𝑥 2 + 2𝑦 2 = 1
7 1 1
Thus the critical points are ± , and 0, ±
8 4 2

𝐿 𝜆; 𝑥, 𝑦 = 𝑥 2 + 𝑦 − 𝜆 𝑥 2 + 2𝑦 2 − 1
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝜆2 𝜕𝜆𝜕𝑥 𝜕𝜆𝜕𝑦
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝐻𝐿 𝜆; 𝑥, 𝑦 =
𝜕𝑥𝜕𝜆 𝜕𝑥 2 𝜕𝑥𝜕𝑦
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝑦𝜕𝜆 𝜕𝑦𝜕𝑥 𝜕𝑦 2
53
 Solution
b) Use the Hessian criterion to determine the nature of the critical
point.
𝐿 𝜆; 𝑥, 𝑦 = 𝑥 2 + 𝑦 − 𝜆 𝑥 2 + 2𝑦 2 − 1
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝜆2 𝜕𝜆𝜕𝑥 𝜕𝜆𝜕𝑦
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝐻𝐿 𝑙; 𝑥, 𝑦 =
𝜕𝑥𝜕𝜆 𝜕𝑥 2 𝜕𝑥𝜕𝑦
𝜕2𝐿 𝜕2𝐿 𝜕2𝐿
𝜕𝑦𝜕𝜆 𝜕𝑦𝜕𝑥 𝜕𝑦 2
0 −2𝑥 −4𝑦
𝐻𝐿 𝑙; 𝑥, 𝑦 = −2𝑥 2 − 2𝜆 0
−4𝑦 0 −4𝜆

54
 Solution
b) Use the Hessian criterion to determine the nature of the critical
point.
0 −2𝑥 −4𝑦
𝐻𝐿 𝜆; 𝑥, 𝑦 = −2𝑥 2 − 2𝜆 0
−4𝑦 0 −4𝜆
(–1)kh2k+1=?
𝑘 = 1, 𝑛 = 2 ⇒ (−1)h3 = −[16𝜆𝑥 2 − 32𝑦 2 + 32𝜆𝑦 2 ]
7 1
At ± , for 𝜆 = 1
8 4

− 16𝜆𝑥 2 − 32𝑦 2 + 32𝜆𝑦 2 = −14 < 0 – local maxima

1
0, ± for 𝜆 = 2/4
2

− 16𝜆𝑥 2 − 32𝑦 2 + 32𝜆𝑦 2 = 16 − 4 2 > 0 – local minima

55
Hessian Criterion for
Constrained Extrema
End of Optional

56
 Applications of Gradient
➢ Directional Derivatives
➢ Critical points
➢ Spherical gradient, Gradient field and
potential
➢ Lagrange Multiplier
➢ Normal Vector

57
Normal Vector

58
 Applications of gradient

➢ Angle between Surfaces

➢ Angle between Line (Vector)
Normal vector
and Surface
to surface
➢ Angle between Curve
and Surface

59
 Level Curves
➢ The level curves of a scalar field Φ(x,y) are
defined by the equation Φ(x,y) = k , where k
is real, constant and in the range of Φ(x,y)

➢ The gradient  Φ(x0,y0) is perpendicular to

the level curve Φ(x,y) = k at the point (x0,y0)

60
 Level Surface
● 2D: Level Curve → 3D: Level Surface
● The level surfaces of a scalar field Φ(x,y,z) are
defined by the equation Φ(x,y,z) = k , where k is
real, constant and in the range of Φ(x,y,z)
● The gradient Φ(x0,y0,z0) is perpendicular to the
level surface Φ(x,y,z) = k at the point (x0,y0,z0)
→ The gradient can be used to construct a normal
vector to a surface at a given point

61
 Normal Vector

In two dimensions, In three dimensions,

𝛻ℎ 𝑥, 𝑦 / 𝛻ℎ 𝑥, 𝑦 is 𝛻𝐹 𝑥, 𝑦, 𝑧 / 𝛻𝐹 𝑥, 𝑦, 𝑧 is
perpendicular to the perpendicular to the
tangential line tangential plane at
at 𝒓 = 𝑥, 𝑦 𝒓 = (𝑥, 𝑦, 𝑧)
62
Normal vector to the surface

63
 Normal vector to surface
Method steps:
1) Convert surface formula 𝑓 𝑥, 𝑦 = 𝑧 into
scalar field
Φ 𝑥, 𝑦, 𝑧 = 𝑓 𝑥, 𝑦 − 𝑧
→ Φ 𝑥, 𝑦, 𝑧 = 0 – level surface

2) Calculate the gradients of the scalar field

3) Insert the point at which the normal vector is to

be calculated into the gradient formula
64
 Normal Vector - example

In 3D, the equation

𝐹 𝒓 = 𝐹 𝑥, 𝑦, 𝑧 = 𝑐
where c is a constant, defines a surface.
Example

2 2 2
𝐹 𝒓 = 𝑥 + 𝑦 + 𝑧 = 𝑐(> 0)

defines a spherical surface with radius 𝑐

𝛻𝐹 𝒓 2𝑥, 2𝑦, 2𝑧 𝒓
ෝ=
𝒏 = =
𝛻𝐹 𝒓 2 𝑥2 + 𝑦2 + 𝑧2 𝑐

65
 Exercise

Find unit normal vector to the curve

𝑥2 − 𝑦2 = 9
at point 𝑥0 , 𝑦0 = (5, −4)

66
 Solution

Find normal vector to the curve

𝑥2 − 𝑦2 = 9
at point 𝑥0 , 𝑦0 = (5, −4)
ℎ 𝑥, 𝑦 = 𝑥 2 − 𝑦 2 = 9
𝛻ℎ 𝑥, 𝑦
𝒏ෝ=
𝛻ℎ 𝑥, 𝑦
2𝑥, −2𝑦 𝑥, −𝑦
= =
2 𝑥2 + 𝑦2 𝑥2 + 𝑦2
ෝ 5, −4 =?
𝒏
5,4
ෝ (5, −4) =
𝒏
41
67
 Exercise

Find normal vector to the surface

𝑥 3 𝑧 + 𝑥 2 𝑦 2 + sin(𝑦𝑧) = 3
at point 𝑥0 , 𝑦0 , 𝑧0 = (−1,0,3)

68
 Solution

Find normal vector to the surface

𝑥 3 𝑧 + 𝑥 2 𝑦 2 + sin(𝑦𝑧) = 3
at point 𝑥0 , 𝑦0 , 𝑧0 = (−1,0,3)

𝛻𝐹 𝑥, 𝑦, 𝑧 𝛻𝐹 𝑥, 𝑦, 𝑧 =?
ෝ=
𝒏
𝛻𝐹 𝑥, 𝑦, 𝑧

𝛻𝐹 𝑥, 𝑦, 𝑧 = 3𝑥 2 𝑧 + 2𝑥𝑦 2 , 2𝑥 2 𝑦 + zcos 𝑦𝑧 , 𝑥 3 + 𝑦 cos(𝑦𝑧)

𝛻𝐹 −1,0,3 =?
ෝ −1,0,3 =?
𝛻𝐹 −1,0,3 = (9,3, −1) 𝒏
9,3, −1
ෝ −1,0,3 =
𝒏
91
69
 Exercise - HW

Find normal vector to the surface

𝑒 𝑥𝑦 + 𝑒 𝑥𝑧 − 2𝑒 𝑦𝑧 = 0
at point 𝑥0 , 𝑦0 , 𝑧0 = (−1, −1, −1)

70
 Solution
Find normal vector to the surface
𝑒 𝑥𝑦 + 𝑒 𝑥𝑧 − 2𝑒 𝑦𝑧 = 0
at point 𝑥0 , 𝑦0 , 𝑧0 = (−1, −1, −1)
𝛻𝐹 𝑥, 𝑦, 𝑧
ෝ=
𝒏
𝛻𝐹 𝑥, 𝑦, 𝑧 𝛻𝐹 𝑥, 𝑦, 𝑧 =?

𝛻𝐹 𝑥, 𝑦, 𝑧 = 𝑦𝑒 𝑥𝑦 + 𝑧𝑒 𝑥𝑧 , 𝑥𝑒 𝑥𝑦 − 2𝑧𝑒 𝑦𝑧 , 𝑥𝑒 𝑥𝑧 − 2𝑦𝑒 𝑦𝑧
𝛻𝐹 −1, −1, −1 =?
𝛻𝐹 −1, −1, −1 = 𝑒(−2,1,1)
ෝ −1, −1, −1 =?
𝒏
𝑒(−2,1,1) (−2,1,1)
ෝ −1, −1, −1 =
𝒏 = 71
𝑒 6 6
 Exercise
Find the normal vector of surface
𝑧 𝑥, 𝑦 = cos 2 𝑥 + sin 𝑥 + cos 𝑦
at a point (𝑥 = 0, 𝑦 = 0 )

72
 Solution
Find the normal vector of surface
𝑧 𝑥, 𝑦 = cos2 𝑥 + sin 𝑥 + cos 𝑦
at a point (x=0,y=0) 𝑧 =? 𝑧 = 2
1) Convert surface formula 𝑓 𝑥, 𝑦 = 𝑧 into scalar field
Φ 𝑥, 𝑦, 𝑧 = 𝑓 𝑥, 𝑦 − 𝑧
→ Φ 𝑥, 𝑦, 𝑧 = 0 – level surface
Φ 𝑥, 𝑦, 𝑧 = cos 2 𝑥 + sin 𝑥 + cos 𝑦 − 𝑧
to which Φ 𝑥, 𝑦, 𝑧 = 0 is a level surface
2) Calculate the gradients of the scalar field
−2cos 𝑥 sin 𝑥 + cos 𝑥
𝛻Φ 𝑥, 𝑦, 𝑧 = −sin 𝑦
−1
The gradient of Ф(x,y,z) is perpendicular to the surface 73
 Solution
The gradient of Ф(x,y,z) is perpendicular to the surface:
−2cos 𝑥 sin 𝑥 + cos 𝑥
𝛻Φ 𝑥, 𝑦, 𝑧 = −sin 𝑦
−1
3) Insert the point at which the normal vector is to be
calculated into the gradient formula

To get the normal vector at a specific point, insert the

coordinates, e.g. x=0 , y=0 1
𝑁1 𝑥 = 0, 𝑦 = 0 = 0
−1
The negative of this vector is −1
also normal to the surface, just 𝑁2 𝑥 = 0, 𝑦 = 0 = 0
pointing in the other direction 1 74
 Exercise
Calculate the angle between
the red and the blue surface at point P:
red: 𝑧 = 2𝑥 2 − 𝑦
1
blue: 𝑧 = −𝑥 4 − 2𝑦 2 + 4 , 𝑃 = 1
1

75
 Solution
Steps:
Calculate the angle between
the red and the blue surface at point P: 1) Convert surface formula
red: 𝑧 = 2𝑥 2 − 𝑦
1 𝑓 𝑥, 𝑦 = 𝑧 into scalar field
blue: 𝑧 = −𝑥 4 − 2𝑦 2 + 4 , 𝑃 = 1
1 Φ 𝑥, 𝑦, 𝑧 = 𝑓 𝑥, 𝑦 − 𝑧
2) Calculate the gradients of the
scalar fields from both surfaces
3) Insert the point at which the
angle is to be calculated into the
gradient formulas
4) Calculate angle by dot-
product
76
 Solution
Step 1: Φ𝑟 = 2𝑥 2 − 𝑦 − 𝑧 , Φ𝑏 = −𝑥 4 − 2𝑦 2 + 4 − 𝑧
Step 2:
4𝑥 −4𝑥 3
𝛻Φ𝑟 = −1 𝛻Φ𝑏 = −4𝑦 Step 3:
−1 −1 P: x= 1 , y=1 , z=1
4 −4
𝐴𝑟 = −1 𝐴𝑏 = −4
−1 −1
Step 4:
Θ −16 + 4 + 1
= arccos
𝐴𝑟 ⋅ 𝐴𝑏 18 33
= arccos −11
𝐴𝑟 𝐴𝑏 = arccos
594
= 116.83𝑑𝑒𝑔 77
 Solution The angle between
Θ = arccos
𝐴𝑟 ⋅𝐴𝑏
= 116.83𝑑𝑒𝑔 surfaces can not be
𝐴𝑟 𝐴𝑏 larger than 90 deg
(unless we define the
4 surfaces to have a
Ԧ𝑏 = 4
𝐴′ Take the negative of
preferential side ! ) –
1 one normal vector
the gradient gives us
one normal vector of
𝐴 𝑟 ⋅ Ԧ𝑏
𝐴′ 16 − 4 − 1 the two possible ones
′
Θ = arccos = arccos
Ԧ𝑏
𝐴𝑟 𝐴′ 18 33
11
= arccos = 63.17𝑑𝑒𝑔
594
Θ′ = 180𝑑𝑒𝑔 − Θ Subtract from
= 180.0 − 116.83 𝑑𝑒𝑔 180 deg to get
= 63.17𝑑𝑒𝑔 actual angle78
Implicit and Explicit Representation of
Surfaces
● Implicit: ● Explicit:

𝑓 𝑥, 𝑦, 𝑧 = 0 𝑓 𝑥, 𝑦 = 𝑧

● suitable to describe ● single valued functions

multivalued functions (but not only (e.g. no sphere)
their differentiation or
integration without dividing
the surface into single-valued
parts)

79
Parametric Representation of Surfaces

➢ Parametrization of a space curve by u:

𝑅 𝑢 = 𝑅1 𝑢 𝑖Ԧ + 𝑅2 𝑢 𝑗Ԧ + 𝑅3 𝑢 𝑘

➢ Analogous to the parametrization of a curve with

one parameter, a surface can be parametrized by
two parameters u and v:

𝑅 𝑢, 𝑣 = 𝑅1 𝑢, 𝑣 𝑖Ԧ + 𝑅2 𝑢, 𝑣 𝑗Ԧ + 𝑅3 𝑢, 𝑣 𝑘
80
 Example
Sphere (radius r): One possible parametrization, many more exist

sin 𝑢 cos 𝑣
0≤𝑢≤π
– parametric: 𝑅 𝑢, 𝑣 = 𝑟 sin 𝑢 sin 𝑣
0 ≤ 𝑣 ≤ 2π
cos 𝑢
– implicit: 𝑥2 + 𝑦2 + 𝑧2 − 𝑟2 = 0

– explicit:
● not possible for the whole sphere
● for the upper half only: 𝑧= 𝑟2 − 𝑥2 − 𝑦2

81
 Example
Paraboloid:

– explicit: 𝑧 = 𝑎𝑥 2 + 𝑏𝑦 2 + 𝑐

– implicit: 𝑎𝑥 2 + 𝑏𝑦 2 + 𝑐 − 𝑧 = 0
𝑢
– parametric: 𝑅 𝑢, 𝑣 = 𝑣
𝑎𝑢2 + 𝑏𝑣 2 + 𝑐
One possible parametrization, many more exist
82
 Normal Vector to Parametrized Surfaces
The partial derivatives with respect to the parameters of the
surface are tangent vectors to the surface:
𝜕𝑅1 𝜕𝑅2 𝜕𝑅3 𝜕𝑅1 𝜕𝑅2 𝜕𝑅3
𝑅𝑢 = 𝑖Ԧ + 𝑗Ԧ + 𝑘, 𝑅𝑣 = 𝑖Ԧ + 𝑗Ԧ + 𝑘
𝜕𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑣 𝜕𝑣 𝜕𝑣
They are base vectors of the tangent plane to the surface
From these two vectors, the normal vector to the surface can be
constructed by the cross-product:
𝑁1 𝑢, 𝑣 = 𝑅𝑢 𝑢, 𝑣 × 𝑅𝑣 𝑢, 𝑣
𝑁2 𝑢, 𝑣 = −𝑁1 𝑢, 𝑣 = 𝑅𝑣 𝑢, 𝑣 × 𝑅𝑢 𝑢, 𝑣
The sign is arbitrary and depends on the parametrization
For closed or convex surfaces, the outer side is usually
regarded as the positive normal vector direction
83
 Exercise
Find the normal vector to the paraboloid parametrized by
𝑣cos 𝑢
0 ≤ 𝑢 ≤ 2π
𝑅 𝑢, 𝑣 = 𝑣sin 𝑢 ,
2
0≤𝑣≤1
𝑣
at point (𝑥 = 1, 𝑦 = 0, 𝑧 = 1)

84
 Solution
Find the normal vector to the paraboloid parametrized by
𝑣cos 𝑢
0 ≤ 𝑢 ≤ 2π
𝑅 𝑢, 𝑣 = 𝑣sin 𝑢 ,
2
0≤𝑣≤1
𝑣
at point (𝑥 = 1, 𝑦 = 0, 𝑧 = 1)
−𝑣sin 𝑢 cos 𝑢
Tangent vectors: 𝑅𝑢 = 𝑅𝑣 = sin 𝑢
𝑣cos 𝑢
0 2𝑣
are base vectors of the tangent plane to the surface. The normal
vector to the surface
𝑖Ԧ 𝑗Ԧ 𝑘 2𝑣 2 cos 𝑢
𝑅𝑢 × 𝑅𝑣= −𝑣sin 𝑢 𝑣cos 𝑢 0 = 2𝑣 2 sin 𝑢
cos 𝑢 sin 𝑢 2𝑣 −𝑣sin2 𝑢 − 𝑣cos2 𝑢
2𝑣 2 cos 𝑢
= 2𝑣 2 sin 𝑢
85
−𝑣
 Solution
Find the normal vector to the paraboloid parametrized by
𝑣cos 𝑢
0 ≤ 𝑢 ≤ 2π
𝑅 𝑢, 𝑣 = 𝑣sin 𝑢 ,
2
0≤𝑣≤1
𝑣
at point (𝑥 = 1, 𝑦 = 0, 𝑧 = 1)

−𝑣sin 𝑢 cos 𝑢
Tangent vectors: 𝑅𝑢 = 𝑣cos 𝑢 , 𝑅𝑣 = sin 𝑢
2𝑣 2 cos 𝑢 0 2𝑣
𝑅𝑢 × 𝑅𝑣 = 2𝑣 2 sin 𝑢
−𝑣
At point (𝑥 = 1, 𝑦 = 0, 𝑧 = 1), → 𝑧 = 1 = 𝑣 2𝑣=1
→ 𝑥 = 1 = 𝑣cos 𝑢 = cos 𝑢 , 𝑢=0
2
𝑁 𝑥 = 1, 𝑦 = 0, 𝑧 = 1 = 𝑁 𝑢 = 0, 𝑣 = 1 = 0
Normal vector at an example point −1 86
 Exercise
Calculating the Angle between a Line
(Vector) and a Surface

What is the (minimum) angle

between the line connecting the line
(grey) and
the surface
x=1 , y=0 , z=1 and intersect at
the point
x=1 , y=1 , z=0 x=1,y=0,z=1

and the surface parametrized

by
𝑣cos 𝑢
0 ≤ 𝑢 ≤ 2π
𝑅 𝑢, 𝑣 = 𝑣sin 𝑢 ,
0≤𝑣≤1
?
2 87
𝑣
 Solution
Angle between Line and Surface
We already have calculated the normal vector to the surface
at the point of intersection (x=1,y=0,z=1):
2
𝑁 𝑥 = 1, 𝑦 = 0, 𝑧 = 1 = 𝑁 𝑢 = 0, 𝑣 = 1 = 0
−1
This is the “outer”
normal vector

The vector corresponding to the line connecting

x=1 , y=0 , z=1 and x=1 , y=1 , z=0 , and pointing away
from the surface is:
1 1 0
𝐴Ԧ = 1 − 0 = 1
88
0 1 −1
 Solution
The normal vector to the surface at the point of intersection (x=1,y=0,z=1):
2
𝑁 𝑥 = 1, 𝑦 = 0, 𝑧 = 1 = 𝑁 𝑢 = 0, 𝑣 = 1 = 0 This is the “outer”
normal vector
−1
The vector corresponding to the line, and pointing away from the surface is:
1 1 0
𝐴Ԧ = 1 − 0 = 1
0 1 −1
The cosine between the vectors is the result of their dot-product
divided by the product of the magnitudes:
𝐴Ԧ ∙ 𝑁 0 2 1 1 If using unit normal
cosφ = = 1 ⋅ 0 = vectors, the cosine is
Ԧ
|𝐴||𝑁| 5 2 10 just the dot-product
−1 −1
1
→ φ = arccos = 71.6𝑑𝑒𝑔
10
The surface is perpendicular to its normal vector
→ angle between line and surface is 89
90 degree minus the result → Θ= 90𝑑𝑒𝑔 − φ = 18.4 𝑑𝑒𝑔