8/6/2020 PRACTICE EXAM IV Maths 21a, O.

Knill, Summer 2020

Name:

• Start by printing your name in the above box.

• Try to answer each question on the same page as the question is asked. If needed, use
the back or the next empty page for work.
• Do not detach pages from this exam packet or unstaple the packet.
• Please try to write neatly. Answers which are illegible for the grader can not be given
credit.
• No notes, books, calculators, computers, or other electronic aids are allowed.
• Problems 1-3 do not require any justifications. For the rest of the problems you have to
show your work. Even correct answers without derivation can not be given credit.
• You have 180 minutes time to complete your work.

1 20

2 10

3 10

4 10

5 10

6 10

7 10

8 10

9 10

10 10

11 10

12 10

13 10

14 10

Total: 150

1
Problem 1) (20 points) No justifications are necessary

1) T F The parametrization ~r(u, v) = [v cos(u), v, v sin(u)] describes a cone.

Solution:
Indeed x2 + z 2 = y 2

The projection vector P~w~ (~v ) of ~v onto w

~ always has length smaller or equal
2) T F than the length of w.
~

Solution:
The projection vector is independent of the length of w.
~

3) T F The vectors ~v = [1, 2, 3] and w

~ = [3, 2, 1] are parallel.

Solution:
Their cross product is not zero.

Let S is the unit sphere and F~ is a vector field in space satisfying div(F~ ) = 0
4) T F
everywhere, then S F~ · dS ~ = 0.
RR

Solution:
By the divergence theorem.

If div(F~ )(x, y, z) = 0 for all (x, y, z) then C F~ · d~r(t) = 0 for any closed
R
5) T F
curve C.

Solution:
The flux integrals though a closed surface would be zero

If F~ (x, y, z) has zero divergence everywhere in space, then F~ has zero curl
6) T F everywhere in space.

2
 Solution:
Get an example, like [−y, x, 0].

If F~ , G
~ are two vector fields which have the same curl then F~ − G
~ is a
7) T F
constant vector field.

Solution:
Take [x, 0, 0], [0, y, 0]. They both have curl zero.

As you know, we also write grad(f ) = ∇f for the gradient. In R3 the

8) T F equation grad(curl(grad(f ))) = ~0 always holds.

Solution:
Since the curl of the gradient is is the zero vector ~0, this reads like grad(~0) = ~0 which
does not make any sense as the gradient is only defined for scalar functions.

9) T F The linearization of f (x, y) = 1 + 3x is the function L(x, y) = 3x.

Solution:
The linearization is 1 + 3x.

If ~r(t), a ≤ t ≤ b is part of a flow line of a vector field F~ (x, y, z) for which

10) T F
|F~ (x, y, z)| = 1 at every point, then | b F~ (~r(t)) dt| is the arc length.
R
a

Solution:
We have F~ (~r(t)) · ~r 0 (t) = ~r (t).

There is a non-constant function f (x, y, z) such that grad(f ) =

11) T F curl(grad(f )).

Solution:
The right hand side is 0.

If F~ is a vector field and E is the solid region x4 + y 4 + z 4 ≤ 1, then

12) T F RRR
~
E div(curl(F )) dxdydz = 0.

3
 Solution:
Because div(curl(F~ ) = ~0.

If the vector field F~ has constant divergence 1 everywhere, then the flux of
13) T F F~ through any closed surface S is the volume of the enclosed solid.

Solution:
By the divergence theorem.

The vector ~j × (~j × ~i) is the zero vector, if ~i = [1, 0, 0], ~j = [0, 1, 0], and
14) T F ~k = [0, 0, 1].

Solution:
~j ×~i = −~k which is perpendicular to ~j. So, the cross product is parallel to ~i and non-zero.

If f (x, y) is maximized at (a, b) under the constraint g = c, then ∇f (a, b)

15) T F and ∇g(a, b) are parallel.

Solution:
These are the Lagrange equations

The distance between a point P and the line L through two different points
16) T F A, B is given by the formula |P~A · AB|/|
~ ~
AB|.

Solution:
One has to use a cross product.

17) T F There exists a vector field F~ such that div(F~ ) = F~ .

Solution:
It is not possible, as the divergence is a scalar.

18) T F The unit tangent vector T~ (t) is perpendicular to the vector T~ 0 (t).

4
 Solution:
We have shown that in class.

The vector field F~ (x, y, z) = [x, 2x, 3x] can not be the curl of an other vector
19) T F
field.

Solution:
Its divergence is not zero

The expression div(grad(div(curl(curl(grad(f )))))) is a well defined function

20) T F of three variables if f is a function of three variables.

Solution:
Every of the expressions make sense. Not well defined would be grad(grad(f )) for example
as the gradient is only defined for scalar functions.

5
Problem 2) (10 points) No justifications are necessary.

a) (2 points) Match the following surfaces. There is an exact match

1 2 3
Parametrized surface ~r(u, v) 1-3
[u, v, exp(−(u2 + v 2 ))]
[sin(v + u), cos(u − v), sin(u)]
[v 2 + u2 , u, v 3 ]

b) (2 points) Match the solids. There is an exact match.

Solid a-c
2 2 2 2 2
x < 8, y + z < 9 (x > 4 or z > 4)
1 ≤ x2 + z 2 ≤ 4, 1 ≤ x2 + y 2 ≤ 4, 1 ≤ y 2 + z 2 ≤ 4
x2 + y 2 + z 2 ≤ 4, x2 < 1, y 2 < 1
a b c
c) (2 points) The figures display vector fields. There is an exact match.

Field I-IV I II III IV

F~ (x, y) = [x, y ]2

F~ (x, y) = [−x, 1]
F~ (x, y) = [x2 , y 2 ]
F~ (x, y) = [y 2 , x]
d) (2 points) Hunt down those partial differential equations!

Equation A-D or 0 if no match Partial differential equations

Clairaut A utt − uxx = 0
Burger B utt + uxx = 0
Wave C ut + uux = 0
Laplace D ut + ux = 0

e) (2 points) In the second midterm, some creative nicknames were created for Fubini: we saw
”Fudini”, ”Fumini”, ”The Italian guy”, ”The French guy”. Some even called him ”Clairaut”.
Lets see again: you task is to chose from the following names ”Euler”, ”Hamilton”, ”Frenet”,
”Clairaut”, ”Fubini”, ”Cauchy”, ”Schwarz”, ”Laplace”, ”Stokes”, ”Gauss”, ”Green”, ”Os-
togradsky”, ”Archimedes”, ”Lagrange”, ”Fermat”.
Name of person or pair of persons
Who showed fxy = fyx ?
Who proved |~v · w|
~ ≤ |~v ||w|?
~
Another joker! We have mentioned the Maxwell equations in the Stokes lecture class.
Here are three of these equations. The fourth one is still missing. If you can write down the
missing one, you can regain 2 points, possibly lost earlier in this problem.

~ = σ,curl(E)
div(E) ~ = −B
~ t , curl(B)
~ =j+E
~t

6
Solution:
2,3,1
b,c,a
II,I,IV,III
O,C,A,B
Caliraut, Cauchy-Schwarz
div(B)=0 (no monopoles).

Problem 3) (10 points) No justifications necessary

a) (5 points) Complete the formulas:

Formula where X is
~v · w
~ = |~v ||w|X
~

|~v × w|
~ = |~v ||w|X
~

P~w~ (~v ) = w(~

~ v · w)/X
~

~u · (~v × w)
~ = 0 if and only if X is zero

If ~v × w
~ = 0, then ~v and w
~ are X

b) (5 points) Choose from the following words to complete the table below:

“arc length formula”, “surface area formula”, “chain rule”, “vol-

ume”, “area of parallelogram”, “partial differential equation”, “Fu-
bini Theorem”, “line integral”, “flux integral”, “Phytagoras theo-
rem”, “Al Khashi cos-formula”, “vector projection”, “scalar projec-
tion”, “partial derivative”, “unit tangent vector”, “normal vector”,
“binormal vector”

7
 Formula Name of formula or rule or theorem

d
dt
f (~r(t)) = ∇f (~r(t)) · ~r 0 (t)

Rb
a |~r 0 (t)| dt

div(curl(F~ )) = 0

~ 2 = |w|
|~v − w| ~ 2 + |~v |2 − 2|~v ||w|
~ cos(θ)

T~ 0 (t)/|T~ 0 (t)|

Solution:
a) cos(θ)
sin(θ)
|w|2
the Volume
parallel.
b) Chain rule
Length
PDE
Al Khashi cos formula
Normal vector

Problem 4) (10 points)

8
A recent viral video features Otmashka Dumin-
ina sky diving from the Aizhai bridge in China.
The clip became famous because she first had to
get rid of police officers who were trying to pre-
vent her jump, wink them good-by and then jump.
There is some wind. The acceleration is

~r 00 (t) = [1 + t, t, −10] .
a) (5 points) Write down ~r(t) and find the po-
sition ~r(2), if ~r(0) = [0, 0, 300] and ~r 0 (0) = [1, 0, 0].

b) (5 points) Find the curvature

|~r 0 (t) × ~r 00 (t)|
|~r 0 (t)|3 Source: Youtube
of the path at ~r(2).

Solution:
a) ~r(t) = [tt2 /2 + t3 /6, t3 /6, −5t2 + 300], ~r(2) = [16/3, 4/3, 280].
b) r0~(2) = [5, 2, −20], ~r00 (2) = [3, 2, −10]. We have ~r0 (2) × ~r00 (2) = [20, −10, 4]. The
curvature is 5161/2 /(429)3/2 .

Problem 5) (10 points)

In this problem, we work with the function f (x, y) = cos(x2 y).

a) (4 points) Estimate f (1.001, π/2 − 0.0001)

using linear approximation.

b) (3 points) Find the tangent plane to the surface

g(x, y, z) = cos(x2 y) − z 2 + 2z = 0 at (1, π/2, 2).

c) (3 points) Find the direction = unit vector

in which the directional derivative D~v f (1, π/2) is
maximal.

9
Solution:
a) ∇f (x, y) = [− sin(x2 y)2xy, sin(x2 y)x2 ]. At the given point this is ∇f (x, y) = [−π, −1].
We estimate 0 − π0.001 + 0.0001.
b) πx + y + √2z = 3π/2 + 4.
c) [−π, −1]/ 1 + π 2 .

Problem 6) (10 points)

While searching for a geometry problem, Oliver sat

at the Harvard square coffee-shop just in front
of spherical lamps located at
A = (1, 2, 4), B = (2, 2, 3), C = (3, 1, 5) .

Well, here it is, the geometry problem:

a) (3 points) Find the area of the triangle ABC.

b) (3 points) Find the equation of the plane

through A,B and C.

b) (4 points) Find the distance of C to the line

through the points A and B. Foto by Oliver taken on August 2, 2016

Solution:
~ × AB|
area |AC
a) The area of the triangle is half of the √ ~ of the parallel epiped spanned
~ and AB.
by the vectors AC ~ The area is 11/2
b) We have the vectors AC ~ × AB ~ = ~n = [2, −1, 1] × [1, 0, −1] = [1, 3, 1]. This gives
the equation x + 3y + z = d. The constant is obtained by plugging in a point. It is
x + 3y + z = 11. √ √
c) The distance is the area of the parallelogram ( 11) divided by the base length 2. It
q
is 11/2 .

Problem 7) (10 points)

10
Find the volume of the solid given by the inequal-
ities
x2 + y 2 ≤ 16
sin(x2 + y 2 ) ≤ z ≤ 2 + cos(x2 + y 2 )
x ≥ 0 or y ≥ 0 .
As the picture illustrates, the last condition means
that the points (x, y, z) for which x < 0, y < 0 are
excluded.

Solution:
After evaluating the most inner integral, we get to the following integral in polar coordi-
nates: Z 5π/2 Z 4
(2 + cos(r2 ) − sin(r2 ))r drdθ .
π 0

This gives 3π/4(31 + sin(16) + cos(16)) . It was also possible to integrate from 0 to 2π
and take 3/4 of the result.

Problem 8) (10 points)

We like it “extreme”. Maximize the function

f (x, y) = x1000 − 1000x + y 200 − 100y 2 .

To do so, find all critical points and classify them
using the second derivative test. This is an inter-
esting example, as computer algebra systems have
trouble writing down the solutions (they look for
solutions in the complex). But we can find the real
ones!

Dean Potter solo walking at Taft Point in Yosemite

11
Solution:
The gradient of f is [1000(x999 − 1), 200y 199 − 200y]. It is zero for x = 1 and y = 0, 1, −1.
There are three solutions. (1, 0) is a saddle, (1, 1), (1, −1) are minima. In order to use
2
the second derivative test, we need D = fxx fyy − fxy . We have fxx = 1000 · 999x998 which
is 1000 · 999 in all three cases. Now, fyy = 200 · 199y 198 − 200. This is −200 for y and
otherwise positive. So, we see that D is negative for (1, 0) and otherwise D > 0. As
fxx > 0 in all the remaining cases, these are all minima.
P.S. What is interesting is that when Mathematica computes the critical points, it get
1000 · 200 critical points in the complex. The output consists of 200’000 entries, this is
one reason why even a computer algebra system starts to stutter, especially if we trained
it to write down a neat table! But we like it extreme!

Problem 9) (10 points)

a) (5 points) On the other extreme side, here is

one of the simplest Lagrange problems: extremize
f (x, y) = x2 − y
on the curve

g(x, y) = y − x − 1 = 0
using the Lagrange method.
b) (5 points) Economists also solve Lagrange
problems by finding the critical points of
F (x, y, z) = f (x, y)−zg(x, y) of three variables. (z
plays the role of the Lagrange multiplier.) Write
down F (x, y, z) for the example given in a) and
find its critical points ∇F (x, y, z) = 0. You should
get the same x, y values as in a).

Solution:
a) The Lagrange equations are 2x = −λ, −1 = λ1 which gives immediately x = 1/2.
Plugging this into the constraint gives y = 3/2. So, the critical point under constraint is
(1/2, 3/2) .
b) The gradient of F (x, y, z) is ∇F (x, y, z) = [fx − zgx , fy − zgy , −g(x, y)]. If we set this
to [0, 0, 0], we which are exactly the Lagrange equations from a) where λ is replaced with
z. Again, we get (1/2, 3/2, −1) .

Problem 10) (10 points)

12
Three days ago, on August 1, 2020, Lukas Irmler
walked over a rope over the Rheinfalls in Switzer-
land. There is a force field F~ present which con-
sists part of the gravitational force and part by the
wind forces:

F~ (x, y, z) = [sin(x), cos(y), −10 + z] .

The path is given by ~r(t) = [5t, t, 30 − sin(t)/10],
where 0 ≤ t ≤ π. Compute the work
Z π
F~ (~r(t)) · ~r 0 (t) dt
0

done by Lukas during this stunt.

Source: Schaffhauser Nachrichten, August 2, 2020

Solution:
Best use the FTLI. The potential is f (x, y, z) = − cos(x) + sin(y) − 10z + z 2 /2. Instead
of integrating, we just have to evaluate f (~r(π)) − f (~r(0)) = f (5π, π, 30) − f (0, 0, 30) = 2.
It was also possible to do the line integral directly, but it was considerably more work.

Problem 11) (10 points)

Find the line integral of the vector field

F~ (x, y) = [−y + x8 , x − y 9 ]
along the boundary C of the generation 4
Pythagoras tree shown in the picture. The curve
C traces each of the 31 square boundaries counter
clockwise. You can use the Pythagoras tree theo-
rem mentioned below. We also included the proof
of that theorem evenso you do not need to read
the proof in order to solve the problem.

13
Pythagoras tree theorem:
The generation n Pythagorean tree has area n + 1.
Proof: in each generation, new squares are added
along a right angle triangle. The 0’th generation is
a square of area c2 = 1. The first generation tree
got two new squares of side length a, b which by
Pythagoras together have area a2 + b2 = c2 = 1.
Now repeat the construction. In generation 2,
we have added 4 new squares which together
have area 1 so that the tree now has area 3. In
generation 3, we have added 8 squares of total
area 1 so that the generation tree has area 4. Etc.
Etc. The picture to the right shows generation 7.
Its area of all its (partly overlapping) leafs is 8.

Solution:
We use the Green theorem. The curl of F~ is constant 2. The integral R curl(F~ ) dxdy
RR

is therefore 2 times the area of R which is 2 × 5 = 10 .

Problem 12) (10 points)

Find the flux of the vector field

F~ (x, y, z) = [4x + z, 2y + x, 3z + x2 + y 2 ]

through the cube octahedron compound E of

volume 12, seen to the right.

P.S. You might be curious to know how the volume

of E is computed: it is the volume of the unit cube
−1 < x < 1, −1 < y < 1, −1 < z < 1 (which is 8)
plus the volumes of 6 pyramids of base area 2 and
height 1 (which gives 4). Together, this is 12.

14
Solution:
We use the divergence theorem. The divergence of the vector field is 9. The integral
~
E div(F ) dV is therefore 9 times the volume of E. The result is 9 · 12 = 108.
RRR

Problem 13) (10 points)

We watch people sailing on the Charles river. A

sail is a surface parametrized by

~r(s, t) = [s sin(t), s(1 − s), t]

0 ≤ s ≤ 1, 0 ≤ t ≤ π. The surface is oriented so
that the two parametrized curves
~r1 (t) = [sin(t), 0, t], 0 ≤ t ≤ π
and
~r2 (t) = [0, 0, π − t], 0 ≤ t ≤ π
together form its oriented boundary. The wind
force is given by
F~ = curl(G)
~ ,G
~ = [z, z sin(z 7 )y, z] .

What is the flux of F~ through the surface S?

Solution:
~ along the two paths.
We use Stokes theorem. This requires to find the line integral of G
We have
Z π Z π
7
[t, t sin(t )0, t] · [cos(t), 0, 1] dt + [t, t sin(t7 )0, π − t] · [0, 0, −1] dt = −2 .
0 0

This simplifies to
Z π Z π Z π Z π
t + t cos(t) dt − t dt = t cos(t) dt = t sin(t)|π0 − sin(t) dt = −2 .
0 0 0 0

In the last step we have used integration by parts. The result is −2 .

Problem 14) (10 points)

15
The vector field

~ y, z) = [−y, x, 0]
A(x,
(x2 + y 2 + z 2 )3/2
is called the vector potential of the magnetic
field
~ = curl(A).
B ~
The picture shows some flow lines of this mag-
~ Find the flux of B
netic dipole field B. ~ through
the lower half sphere x2 + y 2 + z 2 = 1, z ≤ 0 ori-
ented downwards.

Solution:
Since we have an integral of the curl of the vector field A, ~ we use Stokes theorem and
~
integrate A(~r(t)) along the boundary curve ~r(t) = [cos(t), − sin(t), 0]. First of all, we
have A(~~ r(t)) = [sin(t), cos(t), 0]. The velocity is ~r 0 (t) = [− sin(t), cos(t), 0]. The integral
R 2π
is 0 −1 dt = −2π. The answer is −2π .

16