MA2104 Multivariable Calculus Suggested Solutions

AY20/21 Semester 2

Author: Chong Jing Quan, Tan Gian Yion Reviewer: Fang Xinyu

Question 1
Let U⊆R2 be an open set in the plane, and let F : U → R2 be a vector-valued function, with components
u
F= . Suppose F is twice continuously differentiable, and that for all p ∈ U, one has
v
   
0 0 −1 0 −1
F (p) = F0 (p) as a 2 × 2-matrix.
1 0 1 0

(a) Show that

∂2u ∂2u ∂2v ∂2v
2
+ 2 =0= 2
+ 2 on U.
∂x ∂y ∂x ∂y

Write !
∂u ∂u
0 ∂x (p) ∂y (p)
F (p) = ∂v ∂v .
∂x (p) ∂y (p)

Then, we have
! !
∂u ∂u ∂u
− ∂u
  
0 0 −1 ∂x (p) ∂y (p) 0 −1 ∂y (p) ∂x (p)
F (p) = ∂v ∂v = ∂v ∂v
1 0 ∂x (p) ∂y (p)
1 0 ∂y (p) − ∂x (p)

and ! !
∂u ∂u ∂v ∂v
− ∂x (p) − ∂y
   
0 −1 0 −1 ∂x (p) ∂y (p) (p)
F0 (p) = ∂v ∂v = ∂u ∂u .
1 0 1 0 ∂x (p) ∂y (p) ∂x (p) ∂y (p)

This yields
∂u ∂v ∂u ∂v
=− and = .
∂y ∂x ∂x ∂y
Hence, we get

∂2u ∂2u
   
∂ ∂u ∂ ∂u
+ 2 = +
∂x2 ∂y ∂x ∂x ∂y ∂y
   
∂ ∂v ∂ ∂v
= + −
∂x ∂y ∂y ∂x
2 2
∂ v ∂ v
= −
∂x∂y ∂y∂x
= 0 by Clairaut’s Theorem.

Similarly,

∂2v ∂2v
   
∂ ∂v ∂ ∂v
+ = +
∂x2 ∂y 2 ∂x ∂x ∂y ∂y
   
∂ ∂u ∂ ∂u
= − +
∂x ∂y ∂y ∂x
2 2
∂ u ∂ u
=− +
∂x∂y ∂y∂x
= 0.

1
MA2104 Multivariable Calculus AY20/21 Semester 2

(b) Suppose further that for all p ∈ U, one has

F(p) 6= 0 in R2 .

Show that the R-valued function ϕ : U → R defined by

ϕ(p) = log |F(p)| for p ∈ U

satisfies the equation

∂2ϕ ∂2ϕ
+ =0 on U.
∂x2 ∂y 2
(Here, log denotes the natural logarithm function on positive real numbers.)
p
Note that we have |F(p)| = (u(p))2 + (v(p))2 , so log |F(p)| = 12 log((u(p))2 + (v(p))2 ). For conve-
nience, we will simply write u for u(p) and v for v(p).
1
By differentiating 2 log(u2 + v 2 ) twice with respect to x, one has

∂2
   
1 1 ∂ ∂
log(u2 + v 2 ) log(u2 + v 2 )


=
∂x2 2 2 ∂x ∂x
 
1 ∂ 2uux + 2vvx
=
2 ∂x u2 + v 2
1
= ((2u2x + 2uuxx + 2vx2 + 2vvxx )(u2 + v 2 ) − (2uux + 2vvx )2 )
2(u + v 2 )2
2

1
= 2 ((u2x + uuxx + vx2 + vvxx )(u2 + v 2 ) − 2(uux + vvx )2 ). (1)
(u + v 2 )2

Similarly, by differentiating log(u2 + v 2 ) twice with respect to y, one has

∂2 1
 
2 2 1
2
log(u + v ) = 2 ((u2y + uuyy + vy2 + vvyy )(u2 + v 2 ) − 2(uuy + vvy )2 ). (2)
∂y 2 (u + v 2 )2

Using part (a), adding (1) and (2) together yields

1
((u2x + uuxx + vx2 + vvxx )(u2 + v 2 ) − 2(uux + vvx )2 )
(u2 + v 2 )2
1
+ 2 ((u2y + uuyy + vy2 + vvyy )(u2 + v 2 ) − 2(uuy + vvy )2 )
(u + v 2 )2
1
= 2 ((u2x + u2y + vx2 + vy2 )(u2 + v 2 ) − 2(uux + vvx )2 − 2(uuy + vvy )2 )
u + v2
1
= 2 (2(u2x + u2y )(u2 + v 2 ) − 2(uux + vvx )2 − 2(uuy + vvy )2 ). (3)
u + v2
We now observe that

(uux + vvx )2 + (uuy + vvy )2 = (uux )2 + 2uux vvx + (vvx )2 + (uuy )2 + 2uuy vvy + (vvy )2
= (uux )2 − 2uvy vuy + (vvx )2 + (uuy )2 + 2uuy vvy + (vvy )2
= (uux )2 + (vvx )2 + (uuy )2 + (vvy )2
= (u2x + u2y )(u2 + v 2 ). (4)

Substituting (4) into (3) yields the desired result.

2
MA2104 Multivariable Calculus AY20/21 Semester 2

Question 2
(a) Compute the value of the integral Z
(x2 + y 2 )d(x, y)
R

where R is the region in the first quadrant x ≥ 0, y ≥ 0 of R2 bounded by the curves

x2 − y 2 = 1, x2 − y 2 = 4, xy = 1, xy = 3.

Consider the following C 1 transformation

F (x, y) = (x2 − y 2 , xy).

The Jacobian matrix of the transformation is

 
2x −2y
JF = .
y x

It is also easy to see that this is a continuous bijection with continuous inverse from R to the
rectangular domain S = [1, 4] × [1, 3]. This is because we can write the inverse function for F by
solving quadratic equations:
√
s s 
2
a + a + 4b 2 2
F −1 (a, b) =  ,b √ .
2 a + a2 + 4b2

and we see that the function F −1 is also continuously differentiable on the (open) first quadrant.

Thus, by denoting F (x, y) = (a, b), we have det JF = 2x2 + 2y 2 > 0, so det JF−1 = (2x2 + 2y 2 )−1 .
We see that
Z Z Z
1 1
(x2 + y 2 )d(x, y) = (x2 + y 2 )(2x2 + 2y 2 )−1 d(a, b) = d(a, b) = × 6 = 3.
R S S 2 2

(b) Compute the volume of the following subset of R3 :

p
{(x, y, z) ∈ R3 : 2 x2 + y 2 + |z| ≤ 1}.

We first see that by changing to cylindrical coordinates we have

−(1 − 2r) ≤ z ≤ 1 − 2r.

We also see that 0 ≤ r ≤ 21 , and the height h = 2 − 4r. Therefore the volume can be computed using
cylindrical coordinates:
1
Z 2π Z 2
V = (2 − 4r)rdrdθ
0 0
1
Z 2π Z 2
= 2r − 4r2 drdθ
0 0
4r3 21
= 2π × [r2 − ]
3 0
1 π
= 2π × = .
12 6

3
MA2104 Multivariable Calculus AY20/21 Semester 2

Question 3
(a) Show that the function f : R → R defined by
Z 2π
f (ρ) = eρ cos θ cos(ρ sin θ)dθ
0
is a constant function.

First we use the differentiation under the integral sign:

Z 2π
df
= cos θeρ cos θ cos(ρ sin θ) − eρ cos θ sin θ sin(ρ sin θ)dθ. (5)
dρ 0
We will show that this expression evaluates to 0. Assume ρ 6= 0.

Let F~ = (ex sin y, ex cos y) = (P, Q) and let ~r(θ) = (ρ cos θ, ρ sin θ). Let C denote the circle traced by
~r(θ) and R be the region bounded by the circle. Then we see that by the Green’s theorem, we have
Z Z Z
∂Q ∂P
F~ · d~r = − dA = ex cos y − ex cos ydA = 0
C R ∂x ∂y R
On the other hand, we see that by noting x = ρ cos θ and y = ρ sin θ

Z Z 2π
F~ · d~r = F~ · ~r0 (θ) dθ
C 0
Z 2π  x   
e sin y −ρ sin θ
= · dθ
0 ex cos y ρ cos θ
Z 2π
= ρ cos θeρ cos θ cos(ρ sin θ) − ρeρ cos θ sin θ sin(ρ sin θ) dθ
0
df
=ρ .
dρ
df df
Since ρ 6= 0, we have dρ = 0. If ρ = 0, we can also show that dρ = 0 since
Z 2π
df
= cos θdθ = 0.
dρ 0
df
So dρ = 0 on R. Therefore f is a constant function of ρ.
Remark. Observe that the following identity
d ρ cos θ
sin(ρ sin θ) = eρ cos θ ρ cos θ cos(ρ sin θ) − ρ sin θeρ cos θ sin(ρ sin θ)


e
dθ
can be used to compute the integral on (5). Indeed, for ρ 6= 0, one has
Z 2π
df
= cos θeρ cos θ cos(ρ sin θ) − eρ cos θ sin θ sin(ρ sin θ)dθ
dρ 0
Z 2π
1 d ρ cos θ

= e sin(ρ sin θ) dθ
0 ρ dθ
1  ρ cos θ 2π
= e sin(ρ sin θ) 0 = 0.
ρ

Then, one may continue to handle the case for ρ = 0 and finish the proof as above.
(b) Compute the value of the integral
Z 2π
ecos θ cos(sin θ) dθ.
0
The required integral is f (1). Instead, we can compute f (0) since f is constant. So we see that
Z 2π
f (1) = f (0) = 1 dθ = 2π.
0

4
MA2104 Multivariable Calculus AY20/21 Semester 2

Question 4
Let B be the closed unit ball in R3

B = {(x, y, z) ∈ R3 : x2 + y 2 + z 2 ≤ 1}

and let f : R → R be the function given by

f (x, y, z) = xyz.

(a) Determine the global maximum of f on B, as well as the points of B at which f attains this maxi-
mum volume.

First assume that x, y, z ≥ 0. Then we can use the AM-GM inequality:

x2 + y 2 + z 2 2
≥ (xyz) 3
3
Since x2 + y 2 + z 2 ≤ 1, we have xyz ≤ √127 . By letting x = y = z = √13 , we see that the inequality
is tight and the maximum value is indeed √127 . Using the symmetry of the function, we see that the
points of B where f attains maximum are
1 1 1 1 1 1 1 1 1 1 1 1
{( √ , √ , √ ), (− √ , − √ , √ ), (− √ , √ , − √ )( √ , − √ , − √ )}
3 3 3 3 3 3 3 3 3 3 3 3

(b) Compute the surface integral Z

∇f · n dσ
∂B
of the vector field ∇f over the boundary sphere ∂B oriented with the outward pointing unit normal
vectors n.

We use Gauss’s Theorem:

Z Z
∇f · n dσ = ∆f dV
∂B B
∂2f ∂2f ∂2f
Z
= + + dV
∂x2 ∂y 2 ∂z 2
ZB
= 0 dV
B
= 0.

5
MA2104 Multivariable Calculus AY20/21 Semester 2

Question 5
(a) Compute the line integral Z
F · dr
C

of the vector field F : R3 → R3 over the curve C paramaterized by r : [0, 1] → R3 , where

 −y
e − ze−x

πt 1 − e−t
 
log(1 + t)
F(x, y, z) := e−z − xe−y  and r(t) := , sin , .
log(2) 2 1−e
e−x − ye−z

We show that F is conservative by finding a potential function f. Note that one has
∂f
= e−y − ze−x =⇒ f = xe−y + ze−x + h1 (y, z)
∂x
∂f
= e−z − xe−y =⇒ f = ye−z + xe−y + h2 (x, z)
∂y
∂f
= e−x − ye−z =⇒ f = ze−x + ye−z + h3 (x, z)
∂z
for some functions h1 , h2 , h3 . By inspection, one such f is given by

f (x, y, z) = ze−x + ye−z + xe−y .

Since r(0) = (0, 0, 0) and r(1) = (1, 1, 1), it follows from gradient theorem that the integral is
e−1 + e−1 + e−1 − 0 = 3e−1 .
(b) Suppose f : R3 → R is a scalar-valued function and G : R3 → R3 is a vector field, both assumed to
be twice continuously differentiable on R3 . Suppose also that both f and G are of compact support,
which is to say there exists a (solid) ball B in R3 centered at the origin such that f and G are zero
outside B. Show that the dot-product of the gradient vector field ∇f with curl vector field ∇ × G has
a zero volume integral over B, i.e. show that
Z
(∇f ) · (∇ × G) dV = 0.
B

Note that
∇ · (f ∇ × G) = (∇f ) · (∇ × G) + f ∇ · (∇ × G) = (∇f ) · (∇ × G)
since ∇ · (∇ × G) = 0. By Gauss’s Theorem, one has
Z Z
∇ · (f ∇ × G) dV = (f ∇ × G) · n dS = 0
B ∂B

since f and G have compact support and are identically zero on the boundary of B by continuity.
This completes the proof.