1102模模模組組組06-12班班班 微微微積積積分分分3 期期期考考考解解解答答答和和和評評評分分分標標標準準準
1. Suppose we are given a function f(x, y) whose second order partial derivatives are continuous. Fix two points P= (1, −1) and Q = (1.2, −1.1) on the xy-plane. It is known that :
⟨−3, 2, 2⟩ is a normal vector to the surface z = f(x, y) at (1, −1, f(P)),
f(x, y) attains an extreme value at Q.
Answer the following questions.
(a) (2%) Find ∂f
∂x(1, −1) and ∂f
∂y(1, −1).
Solution:
Since(fx(1, −1), fy(1, −1), −1) is normal to the surface z = f(x, y) at (1, −1, f(P)),
we have (fx(1, −1), fy(1, −1), −1) = λ(−3, 2, 2) for some λ (or (fx(1, −1), fy(1, −1), −1) // (−3, 2, 2)). (1%) So λ= −1/2, fx(1, −1) = 3/2 and fy(1, −1) = −1. (1%)
(b) (4%) Use the linearization of f(x, y) at P = (1, −1) to estimate the value of f(1.2, −1.1) − f(1, −1).
Solution:
The linearization of f at (1, −1) is
L(x, y) = f(1, −1) + fx(1, −1)(x − 1) + fy(1, −1)(y + 1) = f(1, −1) +3
2(x − 1) − (y + 1).(2%) Then
f(1.2, −1.1) − f(1, −1) ≈ L(1.2, −1.1) − f(1, −1) =3
2(1.2 − 1) − (−1.1 + 1) = 0.4.(2%)
(c) (1%) (Circle the best answer.) f(Q) is a (i) maximum value (ii) minimum value
Solution:
Answer is (i). (1%)
Since f(Q) is an extreme value and f(Q) > f(P) by (b), we have f(Q) is a maximum value.
(d) (1%) (Circle the best answer.) If fxy(Q) ≠ 0, then fxx(Q) is
(i) positive (ii) non-negative (iii) zero (iv) non-positive (v) negative
Solution:
Answer is (v). (1%)
Since f(Q) is a maximum value, we have fxx(Q) ≤ 0. If fxx(Q) = 0, we have D(Q) = −[fxy(Q)]2< 0 which implies that Q is a saddle point of f . It contradicts to (c) and we have fxx(Q) is negative.
2. Consider the function
f(x, y) =⎧⎪⎪⎪
⎨⎪⎪⎪⎩ xy2
x2+ y2 if(x, y) ≠ (0, 0) 0 if(x, y) = (0, 0) .
(a) (4%) Show that f(x, y) is continuous at (0, 0).
(b) (6%) Find lim
(x,y)→(0,0)
fx(x, y). Is fx(x, y) continuous at (0, 0)? Explain.
(c) (5%) Let u = ⟨a, b⟩ be a unit vector. Use the definition of directional derivatives to find Duf(0, 0). (Express your answer in terms of a and b.)
(d) (3%) Using (c), explain why f(x, y) is not differentiable at (0, 0).
Solution:
(a) Sol 1:
∣f(x, y)∣ = ∣x2y+y22x∣ =x2y+y22∣x∣ ≤ ∣x∣ ≤√ x2+ y2.
Hence ∣f(x, y)∣ → 0 as (x, y) approaches (0,0). (3 pts for showing that lim
(x,y)→(0,0)
f(x, y) = 0) Since lim
(x,y)→(0,0)f(x, y) = 0 = f(0, 0), we conclude that f is continuous at (0,0).
(1 pt for showing that f(x, y) is continuous at (0,0).) Sol 2:
By polar coordinates,
∣f(r cos θ, r sin θ)∣ = ∣r3cos θ sin2θ
r2 ∣ = ∣r∣∣ cos θ∣ sin2θ≤ ∣r∣ → 0 as r → 0.
Hence lim
(x,y)→(0,0)
f(x, y) = limr→0f(r cos θ, r sin θ) = 0 (3 pts for showing that lim
(x,y)→(0,0)
f(x, y) = 0) Since lim
(x,y)→(0,0)f(x, y) = 0 = f(0, 0), we conclude that f is continuous at (0,0).
(1 pt for showing that f(x, y) is continuous at (0,0).) (b) For(x, y) ≠ (0, 0), fx=y2(x(x2+y2+y2)−xy2)22⋅2x= (xy42−x+y22y)22 (2 pts).
fx(0, 0) = limh→0f (h,0)−f (0.0)
h = limh→00−0h = 0 (1 pt) fx(0, y) =yy44 = 1 for all y ≠ 0.
Hence fx(x, y) → 1 ≠ fx(0, 0) as (x, y) approaches (0,0) along the y-axis.
This shows that fx(x, y) is not continuous at (0,0).
(Or we can also show that fx(x, y) → 0 as (x, y) approaches (0,0) along the x-axis. Hence lim
(x,y)→(0,0)fx(x, y) does not exist. fx(x, y) is not continuous at (0,0).)
(3 pts for showing that fx(x, y) is not continuous at (0,0).) (c)
Duf(0, 0) = limh→0f(ah, bh) − f(0, 0)
h (2 pts for definition of Duf(0, 0))
= lim
h→0 h3ab2
h2 − 0
h = lim
h→0ab2= ab2 (3 pts for the answer)
(d) If f(x, y) is differentiable at (0,0), the Duf(0, 0) = afx(0, 0) + bfy(0, 0). (2 pts) However Duf(0, 0) = ab2≠ fx(0, 0)a + fy(0, 0)b.
Therefore we know that f us not differentiable at (0,0). (1 pt)
Page 2 of 8
3. Consider the function I(x, y) = ∫1−yx(t2+ 3t) ⋅ et2dt.
(a) (6%) Find all the critical points of I(x, y).
(b) (6%) Classify the critical points of I(x, y) as local maxima, local minima or saddle points.
Solution:
(a) The partial derivatives are
Ix= (x2+ 3x)ex2,(2 points) Iy= [(1 − y)2+ 3(1 − y)]e(1−y)2.(2 points)
The critical points will be the solutions of Ix= 0 and Iy= 0. Namely, x2+ 3x = 0 and (1 − y)2+ 3(1 − y) = 0 which yield x= 0, −3 and y = 1, 4. So the critical points are (0, 1), (0, 4), (−3, 1), and (−3, 4).(2 points) (b) The critical points we found in (a) all have gradient zero, so we use Second Derivatives Test to classify them.
We compute Ixx, Ixy, and Iyy.
Ixx= (2x + 3)ex2+ (x2+ 3x)ex22x,
Iyy= [2(y − 1) − 3]e(1−y)2+ [(1 − y)2+ 3(1 − y)]e(1−y)22(y − 1).
Both Ixy and Iyx are zero. So
D= IxxIyy− Ixy2 = [2x + 3 + (x2+ 3x)2x]ex2e(1−y)2(2y − 5 + [(1 − y)2+ 3(1 − y)]2(y − 1)).(2 points) D(0, 1) = −9 < 0, so a saddle point. D(0, 4) = 9e9> 0, and Ixx(0, 4) = 3 > 0, so a local min. D(−3, 1) = 9e9> 0, and Ixx(−3, 1) = −3e9< 0, so a local max. Finally, D(−3, 4) = −9e18< 0, a saddle point. In summary, (0, 1) and (−3, 4) are saddle points, (0, 4) is a local min, and (−3, 1) is a local max. (1 point for each critical point.)
4. (12%) By the method of Lagrange multipliers, find the absolute maximum and minimum values of f(x, y, z) = x2− 2y2− 2z2+ 4xz
on the unit sphere x2+ y2+ z2= 1.
Solution:
Let
f(x, y, z) = x2− 2y2− 2z2+ 4xz, g(x, y, z) = x2+ y2+ z2− 1.
By the method of Lagrange multiplier, we have the set of equations
2x+ 4z = 2λx (1)
−4y = 2λy (2)
4x− 4z = 2λz (3)
along with g(x, y, z) = 0. Equation (2) gives either y = 0 or λ = −2. If λ = −2, the system becomes a set of linear equations
3x+ 2z = 0 x= 0
from which we get(x, y, z) = ±(0, 1, 0). If y = 0, we can eliminate λ by x × (3) − z × (1) and obtain 2x2− 3xz − 2z2= (2x + z)(x − 2z) = 0,
hence z=x2,−2x. If z = x2, we have g(x, 0,x2) = 54x2−1 = 0, from which we get (x, y, z) = ± (√25, 0,√1
5). If z = −2x, we have g(x, 0, −2x) = 5x2− 1 = 0, from which we get (x, y, z) = ± (√15, 0,−√25). The values of f at those six critical points are
f(±(0, 1, 0)) = −2, f(±( 2
√5, 0, 1
√5)) = 2, f(±( 1
√5, 0,− 2
√5)) = −3, respectively, so the maximum is 2 and the minimum is−3.
Marking scheme
Setting up the equation of the Lagrange multiplier correctly (4%).
– Incorrect equations with a very minor mistakes (an incorrect sign etc.) (3%).
1% for finding each of the six solutions correctly (6% in total).
– 1% for being aware of the existence of the two cases λ= −2 and y = 0 (with incomplete calculation) – 1% for finiding the fact x= 2z, −z/2 (or the factorization (x − 2z)(2x + z) = 0) for y = 0 case.
Finding correct maximal value (1%), minimal value (1%)
Page 4 of 8
5. Depicted in the Figure, E is an ‘apple-shaped’ solid that, in spherical coordinates, occupies the region E = {(ρ, ϕ, θ) ∶ 0 ≤ ρ ≤ 1 − cos ϕ, 0 ≤ ϕ ≤ π, 0 ≤ θ ≤ 2π}.
It is known that E has a constant density ρ(x, y, z) = 2.
(a) (6%) Find the mass of E.
(b) (8%) Let(0, 0, z) be the center of mass of E. Find z.
Figure. The apple-shaped solid E
Solution:
Prob.5: The following crucial steps must be shown clearly
(a) Total mass:
M = ∫ ∫ ∫E2 dV = 2 ∫02π∫0π∫01−cos ϕρ2sin ϕdρdϕdθ ...2%
= 4π ∫0π(1−cos ϕ)3 3sin ϕ dϕ ...2%
either= 4π3 ∫−11(1 − x)3dx =4π3 [(1−x)4 4]−11 = 16π3
or= 4π3 [(1−cos ϕ)4 4]π0 = 16π3 ...2%
(b) The z-component of center of mass:
¯
z= M1 ∫ ∫ ∫E2z dV =M2 ∫02π∫0π∫01−cos ϕρ3sin ϕ cos ϕ dρdϕdθ ....3%
=Mπ ∫0π(1 − cos ϕ)4sin ϕ cos ϕ dϕ ...2%
either= Mπ ∫−11 x(1 − x)4dx =163[(1−x)5 5 − (1−x)6 6]−11 = −45
or= Mπ[(1−cos ϕ)5 5 − (1−cos ϕ)6 6]π0 = 163
−128
30 = −45 ...3%
6. (a) (9%) Find the volume of the solid that is below the parabolid z= 9 − x2− y2 and above the region enclosed by the lemniscate r2= cos(2θ) on the xy-plane (See Figure).
(b) (9%) Evaluate the triple integral
∭R(36 − x2− 4y2− 9z2) dV where R= {(x, y, z) ∈ R3 ∶ x2+ 4y2+ 9z2≤ 36}.
Figure. The lemniscate r2= cos(2θ)
Solution:
(a) In cylindrical coordinates, the volume is
∫
π/4
−π/4∫
√ cos(2θ)
0 ∫
9−r2 0
r dz dr dθ+ ∫3π/45π/4∫
√ cos(2θ)
0 ∫
9−r2 0
r dz dr dθ By symmetry, the two integrals are the same value.
= 2 ∫−π/4π/4∫
√ cos(2θ)
0 ∫
9−r2 0
r dz dr dθ= 2 ∫−π/4π/4∫
√ cos(2θ) 0
9r− r3 dr dθ
= 2 ∫−π/4π/49 cos(2θ)
2 −cos2(2θ)
4 dθ= ∫−π/4π/49 cos(2θ) −1+ cos(4θ)
4 dθ= 9 −π 8 (b) Let x= 6u, y = 3v, z = 2w, then the triple integral becomes
∭u2+v2+w2≤1(36 − 36x2− 36y2− 36z2)∣J∣ dV = 64∭u2+v2+w2≤1(1 − u2− v2− w2) dV Use spherical coordinates u= ρ sin ϕ cos θ, v = ρ sin ϕ sin θ, w = ρ cos ϕ.
= 64∫
2π
0 ∫
π 0 ∫
1
0 (1 − ρ2)ρ2sin ϕ dρ dϕ dθ= (64)(2π)(2) (1 3−1
5) =3456π 5 Grading: In general, -2% for each big mistake and -1% for each small mistake.
(a) 5% for cylindrical coordinates integral setup, 4% for computation (they get partial credit in computation if their setup is similar to the answer).
(b) 5% for setting up an integral that they can evaluate, 4% for computation (they get partial credit in compu- tation if their setup is similar to the answer).
Note: Both problems can be evaluated using xyz-coordinates.
Page 6 of 8
7. (a) (i) (2%) For each fixed value of y, find ∫
√ 3 1
cos(xy) dx.
(ii) (6%) Use your result in (i) to transform
I= ∫0∞e−y⋅ (sin(√
3y) − sin(y))
y dy
into a double integral and then evaluate I by Fubini’s Theorem.
(b) (10%) Use the change of variables u= xy and v = y to evaluate
∬R y
1+ x2y2dA where R is the region enclosed by the curves xy= 1, xy =√
3, x= 1 and y = 3.
Solution:
(a) (i) Marking scheme for 7(a)(i).
All or nothing. 1M for sign error.
∫
√ 3 1
cos(xy) dx = [sin(xy)
y ]
√ 3
1
= sin(√
3y) − sin(y) y
(ii)
Marking scheme for 7(a)(ii).
(1M) Use (a) to transform I into a double integral (1M) Use Fubini’s theorem (correctly)
(2M)* Correct antiderivative for dy
(1M) For realising limy→∞e−y⋅ (bounded function) = 0 (1M) Correct answer
Remark for *. 1M for any candidates who apply IBP twice but yield incorrect anti-derivative.
∫
∞ 0
e−y⋅ (sin(√
3y) − sin(y))
y dy(a)= ∫0∞∫
√ 3
1 e−y⋅ cos(xy) dx dy
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶
(1M ) Fubini
= ∫
√ 3
1 ∫
∞ 0
e−y⋅ cos(xy) dy dx
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶
(1M )
IBP= ∫
√ 3 1
⎡⎢⎢⎢
⎢⎢⎢⎢
⎢⎣
e−y⋅ (x sin (xy) − cos (xy)) x2+ 1
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶
(2M )
⎤⎥⎥⎥
⎥⎥⎥⎥
⎥⎦
∞
0
dx
= ∫
√ 3 1
1 1+ x2
´¹¹¹¹¹¹¸¹¹¹¹¹¹¶
(1M )
dx
= π
¯12
(1M )
Marking scheme for 7(b).
(2M) correct Jacobian
(1M+1M) correct u- and v-components of the transformed region (okay to just sketch the region)
given region is transformed as a trapezoidal region {(u, v) ∶ 1 ≤ u ≤√
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶3
(1M )
and u≤ v ≤ 3
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶
(1M )
}.
∬R y
1+ x2y2dA= ∫
√ 3
1 ∫
3 u
v 1+ u2 ⋅1
vdv du
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶
(3M )
= ∫
√ 3 1
3− u 1+ u2du
=
⎡⎢⎢⎢
⎢⎢⎢⎢
⎢⎣
3 tan−1(u)
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶
(1M )
−1
2ln(1 + u2)
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¹¶
(1M )
⎤⎥⎥⎥
⎥⎥⎥⎥
⎥⎦
√ 3
1
=π 4 −ln 2
´¹¹¹¹¹¹¹¹¹¹¹¹¹¹¸¹¹¹¹¹¹¹¹¹¹¹¹¹¶2
(1M )
Page 8 of 8