MA1104 Multivariable Calculus AY 2009/2010 Sem 2

NATIONAL UNIVERSITY OF SINGAPORE

MATHEMATICS SOCIETY

PAST YEAR PAPER SOLUTIONS

with credits to Agus Leonardi, Chang Hai Bin, Dr. Ku Cheng Yeaw

MA1104 Multivariable Calculus

AY 2009/2010 Sem 2

Question 1
(a) The equation of tangent plane to the surface f (x, y, z) = k at the point (x0 , y0 , z0 ) is given by
∇f (x0 , y0 , z0 ) · hx − x0 , y − y0 , z − z0 i = 0.

In this case, f (x, y, z) = z 2 − 2x2 − 2y 2 , hence ∇f (x, y, z) = (−4x, −4y, 2z), ∇f (1, −1, 4) =
(−4, 4, 8).

∴ After simplifying, the equation of tangent plane is (x − 1) − (y + 1) − 2(z − 4) = 0.

(b) We have fx = 6y 2 − 6x2 and fy = 12xy − 12y 3 .

For (x, y) to be a critical point, we need fx = fy = 0, or:
 2
y = x2
⇒ (x, y) = (0, 0) or (1, 1) or (1, −1)
y = 0 or x = y 2

Define the discriminant:

D(a, b) = fxx (a, b)fyy (a, b) − [fxy (a, b)]2
We also have
fxx = −12x, fyy = −36y 2 , fxy = 12y
Hence, D(0, 0) = 0, D(1, 1) = 288 > 0 and D(1, −1) = 288 > 0. We also have fxx (1, 1) =
fxx (1, −1) = −12 < 0.

∴ By the second derivative test, (1, −1) and (1, 1) are local maxima.

For the point (0, 0), consider f (x, y) at the plane y = 0. We have f (x, 0) = −2x3 , which does
not have local extremum at x = 0.

∴ (0, 0) is a saddle point.

Hence, there are 3 critical points for the function f (x, y):
the 2 local maxima at (x, y) = (1, −1) and (1, 1),
and the the saddle point at (x, y) = (0, 0)

Question 2
 
f (h, 0) − f (0, 0) 2 1
(a) (i) fx (0, 0) = lim = lim h cos
h→0 h h→0 h3
 
1
For all h 6= 0, we have −h2 ≤ h2 cos ≤ h2 .
h3  
1
As lim −h2 = lim h2 = 0, we have lim h2 cos = 0 by Squeeze Theorem.
h→0 h→0 h→0 h3
Similarly, fy (0, 0) = 0 by symmetry.

NUS Math LaTeXify Proj Team Page: 1 of 6 NUS Mathematics Society

MA1104 Multivariable Calculus AY 2009/2010 Sem 2

(ii) The function is differentiable on (a,b) if there exists functions 1 , 2 , such that:
∆z = f (a + ∆x, b + ∆y) − f (a, b) = fx (a, b)∆x + fy (a, b)∆y + 1 ∆x + 2 ∆y
with 1 , 2 → 0 when ∆x, ∆y → 0.

So, when (a, b) = (0, 0) : f (a, b) = 0, fx (a, b) = 0, fy (a, b) = 0,

If we can find the functions 1 , 2 , such that f (∆x, ∆y) = 1 ∆x+2 ∆y, and lim (1 , 2 ) =
(∆x,∆y)→(0,0)
(0, 0), then the function is differentiable at (0,0).
 
1
f (∆x, ∆y) = (∆x)3 + (∆y)3 · cos


(∆x)3 + (∆y)3
     
2 1 2 1
= (∆x) · cos ∆x + (∆y) · cos ∆y
(∆x)3 + (∆y)3 (∆x)3 + (∆y)3
 
2 1
Since 0 ≤ (∆x) cos ≤ (∆x)2 ,
(∆x)3 + (∆y) 
3
 
2 1
by Squeeze theorem, lim (∆x) cos =0
(∆x,∆y)→(0,0) (∆x)3 + (∆y)3
  
2 1
Similarly, lim (∆y) cos =0
(∆x,∆y)→(0,0) (∆x)3 + (∆y)3
   
2 1
 1 = (∆x) cos


3 3
By letting   (∆x) + (∆y)  , we can conclude that the function is dif-
1
 2 = (∆y)2 cos


(∆x)3 + (∆y)3
ferentiable on (0, 0).
(iii) Note that for all (x, y) 6= (0, 0),
−3x2
    
2 1 3 3

1
fx (x, y) = (3x ) cos + x +y − sin 3
x3 + y 3 x + y3 (x3 + y 3 )2
If we approach (0, 0) along the line x = y, we have:
6t5
   
2 1 1
lim fx (x, y)|(t,t) = lim(3t ) cos + 6 sin
t→0 t→0 2t3 4t 2t3
   
1 3 1
= lim(3t2 ) cos + sin
t→0 2t3 2t 2t3
 
3 1
The limit does not exists, since lim sin does not exists, and
  t→0 2t 2t3
1
lim(3t2 ) cos = 0 (easily shown with Squeeze Theorem).
t→0 2t3
Hence, lim lim fx (x, y) does not exists, and hence fx (x, y) is not continuous at (0, 0).
t→0 (x,y)→(0,0)

(b)
Z 2Z 1 Z 1 Z 2y
sin(y 2 ) dy dx = sin(y 2 ) dx dy
0 x/2 0 0
Z 1
= 2y sin(y 2 ) dy
0
= [− cos(y 2 )]10
= 1 − cos 1

NUS Math LaTeXify Proj Team Page: 2 of 6 NUS Mathematics Society

MA1104 Multivariable Calculus AY 2009/2010 Sem 2

Question 3

(a) All the points in the solid V we are integrating over satisfy:

 0 ≤ x ≤ z,
0 ≤ z ≤ y, ⇐⇒ 0 ≤ x ≤ z ≤ y ≤ a
0 ≤ y ≤ a,


This solid is a tetrahedron (or triangular pyramid), with vertices (x, y, z) = (0, 0, 0), (0, a, 0), (0, a, a),
and (a, a, a).

From the inequality 0 ≤ x ≤ z ≤ y ≤ a, if we integrate dy first, then dz, then dx, the inte-
gration limits will be z ≤ y ≤ a, x ≤ z ≤ a, 0 ≤ x ≤ a.
Z aZ yZ z Z aZ aZ a
f (x) dx dz dy = f (x) dy dz dx
0 0 0
Z0 a Zxa z
= (a − z)f (x) dz dx
0 x
Z a a
z2
= az − f (x) dx
0 2 x
Z a 2
x2

a
= − ax + f (x) dx
0 2 2
1 a
Z
= (a − x)2 f (x) dx
2 0

(b) The solid can be described in cylindrical coordinates as:

√ p
E = {(r, θ, z) : 0 ≤ θ ≤ 2π, 0 ≤ r ≤ 2, 6 − r2 ≤ z ≤ 4 − r2 }

√
Z 2π Z 2 Z 4−r2
Volume = √ r dz dr dθ
0 0 6−r2
√
Z 2π Z 2 p
= 4r − r3 − r 6 − r2 dr dθ
0 0
2π  √2
r4 1
Z
2 2 32
= 2r − + (6 − r ) dθ
0 4 3 0
Z 2π  √

17
= − 2 6 dθ
0 3
√
 
17
= 2π −2 6
3

(c) Using the substitution u = xy, v = x2 y, we have x = u−1 v, y = u2 v −1 . The Jacobian of the
transformation is:
∂(x, y) ∂x ∂y ∂x ∂y
= − = (−u−2 v)(−u2 v −2 − (u−1 )(2uv −1 ) = −v −1
∂(u, v) ∂u ∂v ∂v ∂u

NUS Math LaTeXify Proj Team Page: 3 of 6 NUS Mathematics Society

MA1104 Multivariable Calculus AY 2009/2010 Sem 2

ZZ Z 40 Z 20
xy
e dA = eu | − v −1 | du dv
D 20 10
Z 20 Z 40
1
= eu du dv
10 20 v
= (e20 − e10 ) ln 2

Question 4

(a) Let D = {(L, K) : L ≥ 0, K ≥ 0, 2L + 5K ≤ 150}.

We use the closed interval method to find its absolute maximum.

∂P 400 −1/3 1/3 ∂P 200 2/3 −2/3

First, = L K and = L K . Note that there is no value of K and
∂L 3 ∂K 3
∂P ∂P
L which makes = 0 and = 0.
∂L ∂K
Next, we find the extreme values along the boundary of D.
Along L = 0, we have P (L, K) = 0, hence no critical point there.
Along K = 0, we have P (L, K) = 0, hence no critical point there.
150 − 5K 2/3 1/3
 
Along 2L + 5K = 150, we have P (L, K) = 200 K . To find critical point in this
2
case, we use logarithmic differentiation.
 
2 150 − 5K 1
ln P = ln 200 + ln + ln K
3 2 3
 
dP 1 2
⇔ =P −
dK 3K 3(30 − K)

dP 1 2
Solving = 0, we have P = 0 or = .
dK 3K 3(30 − K)
For the first case, we have K = 0 or 150 − 5K = 0 giving two points (K, L) = (30, 0) and
(K, L) = (0, 75).
For the second case, solving the equation we get another point (K, L) = (10, 50).
Comparing all the values obtained above, we have (K, L) = (10, 50) maximises the production.

(b) Let F(x, y) = hP (x, y), Q(x, y)i. We have

∂P ∂Q
= = −6xy 2
∂y ∂x
Therefore F is conservative.

Moreover, we have ∇f (x, y) = F(x, y) when f (x, y)Z= 2x5 − x2 y 3 .

By fundamental theorem of line integrals, we have F · dr = f (2, 1) − f (0, 0) = 60.
C

(c) Let C be the curve as given in the question, but with parameter domain reduced to 0 ≤ t ≤ π.

NUS Math LaTeXify Proj Team Page: 4 of 6 NUS Mathematics Society

MA1104 Multivariable Calculus AY 2009/2010 Sem 2

The required area, A, is:

I
A=2 x dy
ZCπ
=2 sin t cos2 t dt
0
π
cos3 t

=2 −
3 0
4
=
3

Question 5
(x2 + y 2 + z 2 )3/2 − 3x2 (x2 + y 2 + z 2 )1/2
 
∂ x
(a) We have = .
∂x (x2 + y 2 + z 2 )3/2 (x2 + y 2 + z 2 )3
Proceeding similarly and factoring, we have:
div F = ∇ · F
(x2 + y 2 + z 2 )1/2 3(x2 + y 2 + z 2 ) − 3x2 − 3y 2 − 3z 2
 
=
(x2 + y 2 + z 2 )3
=0

(b) We use the parametric representation

r(φ, θ) = hsin φ cos θ, sin φ sin θ, cos φi, 0 ≤ φ ≤ π, 0 ≤ θ ≤ 2π
We have
rφ × rθ = hsin2 φ cos θ, sin2 φ sin θ, sin φ cos φi
F = hsin φ cos θ, sin φ sin θ, cos φi
Therefore,
F · (rφ × rθ ) = sin φ
The flux is
ZZ ZZ
F · dS = F · (rφ × rθ ) dA
S D
Z 2π Z π
= sin φ dφ dθ
0 0
Z 2π Z π
= dθ sin φ dφ
0 0
= 4π

(c) Consider the region S1 = S 0 − S. By divergence theorem, the flux of F across S1 is equal to
ZZZ
div F dV = 0
S1

The flux of F across S 0 is equal to the sum of that across S and that across S1 , i.e. 4π + 0 = 4π.

NUS Math LaTeXify Proj Team Page: 5 of 6 NUS Mathematics Society

MA1104 Multivariable Calculus AY 2009/2010 Sem 2

Question 6

(a) Assume for a contradiction there exist such a vector field F(x, y, z) = hM (x, y, z), N (x, y, z), P (x, y, z)i.
Since condition (iii) says F and G agree on z = 0, it is necessary that F is not defined at (0, 0, 0).
Moreover, condition (i) - the existence of partial derivatives - implies that F is defined for all
(x, y, z) 6= (0, 0, 0).

Let S be a surface whose boundary is described by x2 + 4y 2 = 1 on the xy-plane with posi-

tive orientation. Moreover, we shall choose S so that it does not contain (0, 0, 0) for the argument
below to work (e.g. we can choose S = {(x, y, z) : x2 + 4y 2 + z 2 = 1, x, y, z ≥ 0} with upward
pointing normal, but not S = {(x, y, 0) : x2 + 4y 2 ≤ 1} since curl F is not defined at (0, 0, 0)).

By condition (ii), ZZ
curl F · dS = 0 (1)
S
On the other hand, we can parametrise the boundary of S by
1
r(t) = hcos t, sin t, 0i, 0 ≤ t ≤ 2π
2
Then
1
r0 (t) = h− sin t, cos t, 0i
2
By Stoke’s Theorem,
ZZ I
curl F · dS = F · dr
S C
Z 2π
= G · r0 (t) dt since F and G agrees on z = 0
0
Z 2π
1 1
= h− sin t, cos t, 0i · h− sin t, cos t, 0i dt
0 2 2
Z 2π
1
= dt
0 2
= π,

contradicting (1). Therefore, no such F exists.

NUS Math LaTeXify Proj Team Page: 6 of 6 NUS Mathematics Society