THE UNIVERSITY OF HONG KONG

DEPARTMENT OF MATHEMATICS

MATH2211 Multivariable Calculus

Test 2 Solutions

1. (a) 0.44

The approximation is given by

f (1, 1) + fx (1, 1)(0.9 − 1) + fy (1, 1)(0.8 − 1)

1
+ (fxx (1, 1)(0.9 − 1)2 + 2fxy (1, 1)(0.9 − 1)(0.8 − 1) + fyy (1, 1)(0.8 − 1)2 )
2
1
= 1 + 2(−0.1) + 3(−0.2) + (4(−0.1)2 + 10(−0.1)(−0.2) + 6(−0.2)2 )
2
= 0.44.

1 5
(b) (e − e2 − e3 + 1)
6
Note that e2x+3y > 0. By proposition 4.1 and Fubini’s theorem, the volume is given by
1 1
∬ e2x+3y dA = ∫ ∫ e2x+3y dy dx
[0,1]×[0,1] 0 0
1 y=1
1
=∫ [ e2x+3y ] dx
0 3 y=0
1 1
= ∫ (e2x+3 − e2x ) dx
3 0
x=1
1 1 1
= [ e2x+3 − e2x ]
3 2 2 x=0
1 5
= (e − e − e + 1).
2 3
6

2. (a) True. Let g(x, y, z) = x4 + y 4 + 2z 2 . Note that

∇g(x, y, z) = (4x3 , 4y 3 , 4z) ≠ (0, 0, 0)

since (0, 0, 0) does not satisfy g(x, y, z) = A. Also, ∇f (x, y, z) = (2xy + z, x2 + z, x + y). By
the Lagrange multiplier method, there exists λ such that
⎧
⎪ 2ab + c = 4λa3 , (1)
⎪
⎪
⎨ a2 + c = 4λb3 ,
⎪
⎪
⎪
⎩ a + b = 4λc. (2)

Using (1) and (2), we obtain

2abc + c2 = c(2ab + c) = 4λa3 c = a3 (a + b) = a4 + a3 b.

1
 √ √
(b) False. The bounds −2 ⩽ x ⩽ 2 and
√ − 4 − x ⩽ y ⩽ 4 − x correspond
2 2
√ to the disc x2 +y 2 ⩽ 4.
However, since z ⩾ 1 and z ⩽ 4 − x2 − y 2 , we see that 1 ⩽ 4 − x2 − y 2 . This means the
correct bounds for x and y are determined by the disc x2 + y 2 ⩽ 3 instead. This shows the
bounds of the iterated integral are incorrect. For example, if f (x, y, z) = 1, the volume
described by the iterated integral is different from that described by the triple integral.
So the results are not the same.

3. We solve ∇f (x, y, z) = (4x + 2z − 10, 6y − 18, 4z + 2x − 8) = (0, 0, 0), i.e.

⎧
⎪ 4x + 2z − 10 = 0, (1)
⎪
⎪
⎨ 6y − 18 = 0, (2)
⎪
⎪
⎪
⎩ 4z + 2x − 8 = 0. (3)

Equation (2) gives y = 3. Equation (1) gives z = 5 − 2x. Putting this into (3), we obtain
12 − 6x = 0. So x = 2 and z = 1. We check that (2, 3, 1) satisfy all equations. So it is the only
critical point of f . Next, we compute
x y z
⎛4 0 2⎞
Hf (x, y, z) = ⎜0 6 0⎟ .
⎝2 0 4⎠

RRR4 0 2RRRR
R
∣ = 24 > 0 and RRRRR0 6 0RRRR = 72 > 0
4 0
The leading principal minors of Hf (2, 3, 1) are 4 > 0, ∣ R
0 6 RRR2 0 4RRRR
R
respectively. By Sylvester’s criterion, Hf (2, 3, 1) is positive definite. Therefore, (2, 3, 1) is a
point of local minimum.

4. The global extrema exist by the extreme value theorem since f is continuous and the triangular
region is compact.

The region is determined by the constraints x ⩾ 0, y ⩾ 0 and x + y ⩽ 4. For interior

−(x − 1)2
points, we solve ∇f (x, y) = (0, 0). From = 0, we have x = 1. This also satis-
(x + y + 1)2

2
 (x − 1)(x + 2y + 3)
fies = 0. Therefore, all points (1, y) with 0 < y < 3 are interior critical
(x + y + 1)2
3y + 6
points, where f (1, y) = = 3.
y+2

(0, 4)

x
(0, 0) (4, 0)

Next, we consider the boundary points.

3y + 4 1
ˆ For x = 0 and 0 ⩽ y ⩽ 4, f (0, y) = = 3+ . This is decreasing in y. Thus, the
y+1 y+1
16
extrema can be f (0, 0) = 4 and f (0, 4) = .
5
x2 + x + 4 4 4
ˆ For y = 0 and 0 ⩽ x ⩽ 4, f (x, 0) = = x+ . Since its derivative is 1− ,
x+1 x+1 (x + 1)2
it is decreasing on [0, 1] and increasing on [1, 4]. Thus, the extrema can be f (0, 0) = 4,
24
f (1, 0) = 3 and f (4, 0) = .
5
x2 − 2x + 16 (x − 1)2 + 15
ˆ For y = 4 − x and 0 ⩽ x ⩽ 4, f (x, 4 − x) = = . Its graph is a
5 5
16 24
parabola. Thus, the extrema can be f (0, 4) = , f (1, 3) = 3 and f (4, 0) = .
5 5
24
By comparing the above values, we find that the global maximum is and the global
5
minimum is 3.

5. Using polar coordinates, the condition 1 ⩽ x2 + y 2 ⩽ 4 becomes 1 ⩽ r ⩽ 2 and 0 ⩽ θ ⩽ 2π.

Therefore, by theorem 4.3,
2π 2
∬ (x + y − 2x) dx dy = ∫ (r2 − 2r cos θ)2 r dr dθ
2 2 2
∫
D 0 1
2π 2
=∫ ∫ (r5 − 4r4 cos θ + 4r3 cos2 θ) dr dθ
0 1
2π r=2
1 4
=∫ [ r6 − r5 cos θ + r4 cos2 θ] dθ
0 6 5 r=1
2π 21 124 15(cos 2θ + 1)
=∫ ( − cos θ + ) dθ
0 2 5 2
θ=2π
124 15
= [18θ − sin θ + sin 2θ]
5 4 θ=0
= 36π.