SF1686 Multivariable calculus

Solutions to exam 23.12.18

D EL A

(1) The inequalities 0 ≤ y ≤ x and 1 ≤ x ≤ 2 define the compact set D2 , and we let D
denote the set defined by 0Z≤
Z y ≤ x and 1 ≤ x.
dA
(a) Compute the integral 3
. (4)
D2 (x + 2y) ZZ
dA
(b) Determine the value of the generalised integral 3
. (2)
D (x + 2y)
Lösning. Let Dn be the compact set 0 ≤ y ≤ x and where 1 ≤ x ≤ n, for some integer
n > 1. We then have that
Z nZ x Z 1
 −1
Z
dA dy dx −2 x

3
= 3
= (x + 2y) 0
.
Dn (x + 2y) 1 0 (x + 2y) 0 2·2
We have that (3x)−2 − x−2 = − 98 x−2 , and the above integral reads
Z 1
2 −2 2 n 2 1
x dx = − x−1 1 = (1 − ).
0 9 9 9 n
dA 1
RR
a) With n = 2 we get that D2 (x+2y)3 = 9 .
b) We have that
ZZ ZZ
dA dA 2 1 2
3
= lim 3
= lim (1 − ) = .
D (x + 2y) n→∞ Dn (x + 2y) n→∞ 9 n 9
□
2 SF1686 – Lösningsförslag till tentamen 23.12.18

(2) The function r(t) = (3 cos t, 3 sin t, 4t), with t ≥ 0, describes the position of a particle at
a given time t.
(a) Determine the velocity (or speed) vector v(t) of the particle at time t. (2)
(b) Describe and draw the curve determined by the velocity v(t). (2)
(c) Compute the arc length of the curve r(t) between 1 ≤ t ≤ 2. (2)
Lösning. a) We have that v(t) = (−3 sin(t), 3 cos(t), 4).
R 2 by v(t) isR a2 √
b) The curve determined circle with radius 3, lying in the plane z = 4.
c) The arc length is 1 |v(t)| dt = 1 9 + 16 dt = 5. □

D EL B

(3) We have the function f (x, y) = x2 + y 2 + x2 y + 4 defined for all pairs x, y.

(a) Determine the critical points of f . (2)
(b) Determine the types of critical points of f . (2)
(c) Determine if f has a minimum value. (2)
Lösning. a) The critical points are solutions to ∇f = (2x + 2xy, 2y + x2 = (0, 0). The
equation 2x(1 + y) = 0 gives that either x = 0 or y = −1. Plugging in these values
into√the equation 2y + x2 = 0 gives us the following three critical points, (0, 0) and
(± 2, −1).
b) We apply the Hessian test to determine the types of critical points. We have that
f1,1 = 2 + 2y, that f1,2 = 2x, and that f2,2 = 2. Evaluating these expression at (0, 0)
2
gives that f1,2 − f1,1 f2,2 = −4 < 0, and since f1,1 > 0 we have that the function has a
local minimum at (0, 0). √ 2
Evaluation at the two remaining critical points (± 2, −1) gives that f1,2 − f1,1 f2,2 =
√
8 > 0, and consequenlty the function has a saddle point at (± 2, −1).
c) No, the function does not have a minimum value. We have that x2 +x2 y = x2 (1+y)
is negative when y < −1. For any positive c > 0 we have that
f (x, −1 − c) = x2 (1 + (−1 − c) + c2 + 4 = −cx2 + c2 + 4,
and this value diverges towards − when x grows. □
 SF1686 – Lösningsförslag till tentamen 23.12.18 3

(4) The function φ(x, y) = arctan(y/x) is defined on V = {(x, y) | x > 0}. The vector
1
field F = x2 +y 2 (−y, x) is defined on U = {(x, y) | (x, y) ̸= (0, 0)}. The curve C starts

in (1, −1) and follows the ellipse 4(x − 1)2 + y 2 = 1 counter clockwise to the point
(1, −1).
d 1
(a) Determine the gradient fieldR of φ. (Recall that dt arctan(t) = 1+t2 ). (2)
(b) Compute the line integral C F · dr. (2)
(c) Is the vector field F conservative over U? (2)
Lösning. a) The gradient field of φ is the vector field F , defined on the smallerR set V .
b) The curve C is in V where the field is conservative. Therefore the integral C F·dr =
0.
b) Alternative. There was a typo in the text, and the intended problem was to compute
the line integral of the curve C starting in (1, −1) and ending in (1, 1). In this case the
curve C is in V , so
Z
π
F · dr = φ(1, 1) − φ(1, −1) = .
C 2
c) No, the vector field is not conservative.
R For instance, if we let C denote a circle
centered around (0, 0), then we have that C F · dr = 2π. The curve is closed, the curve
integral of C through F is not zero, and therefore the vector field is not conservative. □
4 SF1686 – Lösningsförslag till tentamen 23.12.18

D EL C

(5) We have the vector field F(x, y, z) = (x2 y, y + z, x2 − y). Compute the flux of F up
(positive z-direction) through the surface S definied by y 2 + z 2 = 4, where 0 ≤ x ≤ 2
and z ≥ 0. (6)

Lösning. PARAMETRISATION
The surface S we parametrise as r(s, t) = (s, 2 cos t, 2 sin t) where D is the rectangle
0 ≤ s ≤ 2, 0 ≤ t ≤ π. This particular parametrisation gives us

F = (2s2 cos t, 2 cos t + 2 sin t, s2 − 2 cos t)

n̂ dS = ±(r′s × r′t ) ds dt = ±(0, −2 cos t, −2 sin t) ds dt

where ± is replaced with minus sign in order to have the correct orientation of the normal.
Then the flux is computed as
ZZ ZZ
32
F · n̂ dS = (4 cos2 t + 2s2 sin t) ds dt = + 4π.
S D 3

D IVERGENCE T HEOREM

An alternative solution is to use the Divergence Theorem. The surface S is not the
boundary of a solid, but we can add the surfaces

S1 = {(x, y, z) : x = 2, y 2 + z 2 ≤ 4, z ≥ 0},
S2 = {(x, y, z) : x = 0, y 2 + z 2 ≤ 4, z ≥ 0},
S3 = {(x, y, z) : 0 ≤ x ≤ 2, −2 ≤ y ≤ 2, z = 0}

in order to have a boundary of a solid K. The Divergence Theorem then gives the identity
ZZ ZZZ
F · n̂ dS = div F dV.
S+S1 +S2 +S3 K

The left term is the sum of four flux integrals, and we obtain that the sought flux through
the surface S is
ZZZ ZZ ZZ ZZ
div F dV − F · n̂ dS − F · n̂ dS − F · n̂ dS.
K S1 S2 S3
 SF1686 – Lösningsförslag till tentamen 23.12.18 5

On S1 , S2 and S3 the outwards pointed unit normal field is n̂ = (1, 0, 0), n̂ = (−1, 0, 0)
and n̂ = (0, 0, −1), respectively. We have that
ZZZ ZZ ZZ
div F dV = (2xy + 1) dV = 1 dV = vol K = 4π,
K K K
ZZ ZZ
F · n̂ dS = 2 2
4y dx dy = {symmetry} = 0,
S1 y +z ≤4
z≥0
ZZ ZZ
F · n̂ dS = 0 dx dy = 0,
S2 y 2 +z 2 ≤4
z≥0
ZZ ZZ
32
F · n̂ dS = (y − x2 ) dx dy = − .
S3
0≤x≤2 3
−2≤y≤2

The flux through S is therefore 4π + 32/3. □

6 SF1686 – Lösningsförslag till tentamen 23.12.18

2 2
(6) To each point P on the ellipse x8 + y18 = 1 we let γP be the line segment from origin to
P . Determine the points on the ellipse that maximise the curve integral
Z
I(P ) = F • dr,
γP

with the vector field F(x, y) = (y, x). (6)

Lösning. The line segment γP is parametrised by r(t) = (x(t), y(t)) = (tx, ty) , with
0 ≤ t ≤ 1, and where the point P = (x, y). The velocity is
dr
= (x, y) .
dt
We get that
dr
F• = (y(t), x(t)) • (x, y) = 2txy.
dt
With the given parametrisation we have that the curve integral becomes
Z Z 1
I(x, y) = I(P ) = F • dr = 2txydt = xy
γ(P ) 0

Therefore, the problem is to maximise the value of the continuous function I(x, y) on
R also have realised by using the potential φ(x, y) = xy of
the ellipse. (This we could
the vector field F. Then γP F · dr = φ(P ) − φ(0, 0) = xy.) The ellipse is compact,
and therefore we have that there exists extrem values. We use the Lagrange multiplicator
method to find these values,
 2
y2

x
L(λ, x, y) = I(x, y) + λg(x, y) = xy + λ + −1 .
8 18
We search for critical points of the function L,
∂L x
0= =y+λ·
∂x 4
∂L y
0= =x+λ·
∂y 9
2 2
∂L x y
0= = + −1
∂λ 8 18
The two first equations are used to eliminate λ, and we have
y2 x2
= .
18 8
This identity is used in the third equation and we obtain x = ±2. And therefore that
y = ±3. The points (2, 3) and (−2, −3) are such that I = Imax = +6. The points (−2, 3)
and (2, −3) are such that I = Imin = −6.
The conclusion is that I(x, y) obtains its maximum at (2, 3) and at (−2, −3), and the
value isI = Imax = +6. □