Calculus 0413 Midterm 1

National Central University, Mar. 29, 2021

Problem 1. (20%) True or False:

In the following, R is always an open region in the plane, (a, b) is always a point in R, and
f : R Ñ R is a function of two variables.

F 1. If lim f (a + t cos θ, b + t sin θ) exists for all θ P R, then lim f (x, y) exists.
tÑ0 (x,y)Ñ(a,b)

F 3. If fx and fy both exist on R, then f is differentiable on R.

T 5. If fx and fy both exist and are bounded on R, then f is continuous on R.

F 7. If the directional derivative of f at (a, b) exists in all directions, then f is continuous at (a, b).

F 9. If fx and fy are continuous on R, then the level curve f (x, y) = f (a, b) has a tangent line at
(a, b).

Problem 2. Let R be an open region in the plane, f : R Ñ R be a function, and (a, b) P R.

1. (5%) Define the differentiability of f at (a, b).

2. (5%) Define the directional derivative of f at (a, b) in direction u, where u = (cos θ, sin θ) is a
unit vector.

Problem 3. Assume that f is a continuous function of two variable satisfying that

f (x, y) ´ 3x2 + 2y 2
lim a = 0.
(x,y)Ñ(´1,1) (x + 1)2 + (y ´ 1)2

1. (10%) Find fx (´1, 1) and fy (´1, 1).

2. (5%) Prove or disprove that f is differentiable at (´1, 1).

f (x, y) ´ 3x2 + 2y 2
Solution. Note that since lim a = 0, we must have
(x,y)Ñ(´1,1) (x + 1)2 + (y ´ 1)2
[ ]
lim f (x, y) ´ 3x2 + 2y 2 = 0 ;
(x,y)Ñ(´1,1)

thus lim f (x, y) = 1. Since f is continuous, f (´1, 1) = 1.

(x,y)Ñ(´1,1)
For (x, y) ‰ (´1, 1),
[ ]2 [ ]2
f (x, y) ´ 3x2 + 2y 2 f (x, y) ´ 3 (x + 1) ´ 1 + 2 (y ´ 1) + 1
a = a
(x + 1)2 + (y ´ 1)2 (x + 1)2 + (y ´ 1)2
f (x, y) ´ 3(x + 1)2 + 6(x + 1) ´ 3 + 2(y ´ 1)2 + 4(y ´ 1) + 2
= a
(x + 1)2 + (y ´ 1)2
f (x, y) ´ f (´1, 1) + 6(x + 1) + 4(y ´ 1) 3(x + 1)2 + 2(y ´ 1)2
= a +a .
(x + 1)2 + (y ´ 1)2 (x + 1)2 + (y ´ 1)2
 ˇ 3(x + 1)2 + 2(y ´ 1)2 ˇ
ˇ ˇ
Since ˇ a 2 2
ˇ ď 3|x + 1| + 2|y ´ 1|, by Squeeze Theorem we find that
(x + 1) + (y ´ 1)

3(x + 1)2 + 2(y ´ 1)2

lim a = 0.
(x,y)Ñ(´1,1) (x + 1)2 + (y ´ 1)2

Therefore,
f (x, y) ´ f (´1, 1) + 6(x + 1) + 4(y ´ 1)
lim a =0
(x,y)Ñ(´1,1) (x + 1)2 + (y ´ 1)2
which implies that
ˇ ˇ
ˇf (x, y) ´ f (´1, 1) + 6(x + 1) + 4(y ´ 1)ˇ
lim a = 0.
(x,y)Ñ(´1,1) (x + 1)2 + (y ´ 1)2

1. Note that the identity above implies that

ˇ ˇ
ˇf (x, y) ´ f (´1, 1) + 6(x + 1) + 4(y ´ 1)ˇ
lim a = 0.
(x,y)Ñ(´1,1)
(x + 1) 2 + (y ´ 1)2
y=1

Therefore,
ˇ ˇ
ˇf (x, y) ´ f (´1, 1) + 6(x + 1) + 4(y ´ 1)ˇ ˇˇ
0= lim a ˇ
(x,y)Ñ(´1,1)
y=1
(x + 1)2 + (y ´ 1)2
ˇ f (x, 1) ´ f (´1, 1) + 6(x + 1) ˇ ˇ f (x, 1) ´ f (´1, 1) ˇ
= lim ˇ ˇ = lim ˇ + 6ˇ ;
ˇ ˇ ˇ ˇ
xÑ´1 x+1 xÑ´1 x ´ (´1)

thus
f (x, 1) ´ f (´1, 1)
fx (´1, 1) = lim = ´6 .
xÑ´1 x ´ (´1)
Similarly, fy (´1, 1) = ´4.

2. In the computations above, we conclude that

ˇ ˇ
ˇf (x, y) ´ f (´1, 1) ´ fx (´1, 1)(x + 1) ´ fy (´1, 1)(y ´ 1)ˇ ˇˇ
lim a ˇ = 0.
(x,y)Ñ(´1,1) (x + 1)2 + (y ´ 1)2

By definition, f is differentiable at (´1, 1). ˝

Problem 4. (10%) Let f, g : R2 Ñ R be defined by

$ 2
& x (x + y) if (x, y) ‰ (0, 0) ,
f (x, y) = x2 + y 4
0 if (x, y) = (0, 0) .
%

Find the directional derivative of f at (0, 0) in the direction along which the value of the function f
at (0, 0) decreases most rapidly.
Solution. Let u be the direction along which the value of the function f at (0, 0) decreases most
rapidly. Then
␣ ˇ (
(Du f )(0, 0) = min (Dv f )(0, 0) ˇ }v} = 1 .

Let v = (cos θ, sin θ). Then

f (t cos θ, t sin θ) ´ f (0, 0) t3 cos2 θ(cos θ + sin θ)

(Dv f )(0, 0) = lim = lim 3
tÑ0 t tÑ0 t (cos2 θ + t2 sin4 θ)

cos2 θ(cos θ + sin θ)

= lim .
tÑ0 cos2 θ + t2 sin4 θ

If cos θ = 0, then (Dv f )(0, 0) = 0. If cos θ ‰ 0, then (Dv f )(0, 0) = cos θ + sin θ. Therefore,

if cos θ = 0 ,
"
0
(Dv f )(0, 0) =
cos θ + sin θ if cos θ ‰ 0 .
␣ ˇ ( ? ( 3π ) ?
Since min cos θ + sin θ ˇ θ P [0, 2π) = ´ 2 attained at θ = ; thus (Du f )(0, 0) = ´ 2. ˝
4
y+1
Problem 5. (15%) Find the second Taylor polynomial of the function f (x, y) = arctan at
x+1
(0, 0).
π
Solution. First, f (0, 0) = arctan 1 = . By the chain rule, for x ‰ ´1,
4
B y+1 y+1
B x x+1
´ (x+1)2 y+1
fx (x, y) = y+1 2 = y+1 2 =´ ,
1 + ( x+1 ) 1 + ( x+1 ) (x + 1)2 + (y + 1)2
B y+1 1
B y x+1 x+1 x+1
fy (x, y) = y+1 2 = y+1 2 = ,
1 + ( x+1 ) 1 + ( x+1 ) (x + 1)2 + (y + 1)2

and
2(x + 1)(y + 1) ´2(x + 1)(y + 1)
fxx (x, y) = [ ]2 ,
fyy (x, y) = [ ]2
(x + 1)2 + (y + 1)2 (x + 1)2 + (y + 1)2
(x + 1)2 + (y + 1)2 ´ 2(y + 1)2 (y + 1)2 ´ (x + 1)2
fxy (x, y) = ´ [ ]2 =[ ]2 .
(x + 1)2 + (y + 1)2 (x + 1)2 + (y + 1)2

Therefore, the second Taylor’s polynomial of f is

1[ ]
f (0, 0) + fx (0, 0)x + fy (0, 0)y + fxx (0, 0)x2 + 2fxy (0, 0)xy + fyy (0, 0)y 2
2
π 1 1 1 ( 1 2 1 2) π 1 1 1( )
= ´ x+ y+ x ´ y = ´ x + y + x2 ´ y 2 . ˝
4 2 2 2 2 2 4 2 2 4
2 ´x2
Problem 6. (10%) Find all relative extrema and saddle points of f (x, y) = (x2 + y 2 )ey using the
second derivative test. When a relative extremum is found, determine if it is a relative maximum or
a relative minimum.
Solution. We first compute the first and second partial derivatives of f and find that
2 ´x2 2 ´x2 2 ´x2
fx (x, y) = 2xey + (x2 + y 2 )(´2x)ey = 2x(1 ´ x2 ´ y 2 )ey ,
y 2 ´x2 y 2 ´x2 y 2 ´x2
fy (x, y) = 2ye + (x2 + y 2 )(2y)e = 2y(1 + x2 + y 2 )e ,
[ ] 2 2
fxx (x, y) = 2 ´ 6x2 ´ 2y 2 ´ 4x2 (1 ´ x2 ´ y 2 ) ey ´x ,
[ ] 2 2
fxy (x, y) = 2x(´2y) + 4xy(1 ´ x2 ´ y 2 ) ey ´x ,
[ ] 2 2
fyy (x, y) = 2 + 2x2 + 6y 2 + 4y 2 (1 + x2 + y 2 ) ey ´x .

Therefore, critical points of f are (0, 0), (1, 0) and (´1, 0).
1. Since fxx (0, 0) = fyy (0, 0) = 2, fxy (0, 0) = 0, we find that fxx (0, 0)fyy (0, 0) ´ fxy (0, 0)2 = 4 ą 0;
thus the fact that fxx (0, 0) ą 0 implies that f (0, 0) is a relative minimum of f .

2. Since fxx (1, 0) = ´4e´1 , fyy (1, 0) = 4e´1 and fxy (1, 0) = 0, we find that fxx (0, 0)fyy (0, 0) ´
fxy (0, 0)2 = ´16e´2 ă 0; thus (1, 0) is a saddle point of f .

3. Since fxx (´1, 0) = ´4e´1 , fyy (´1, 0) = 4e´1 and fxy (´1, 0) = 0, we find that fxx (0, 0)fyy (0, 0)´
fxy (0, 0)2 = ´16e´2 ă 0; thus (´1, 0) is a saddle point of f . ˝
Problem 7. (20%) Let R be the solid in the space given by
␣ ˇ a (
(x, y, z) ˇ 1 ď z ď 4 ´ x2 ´ y 2 .

Find the extreme value of function w = f (x, y, z) = xyz on R.

Solution. Let g(x, y, z) = x2 + y 2 + z 2 ´ 4, and h(x, y, z) = z ´ 1. Then

(∇f )(x, y, z) = (yz, xz, xy) ,

(∇g)(x, y, z) = (2x, 2y, 2z) ,
(∇h)(x, y, z) = (0, 0, 1) .

If (∇f )(x, y, z) = 0, then xy = yz = zx = 0 which implies that at least two of x, y, z are zero. In
this case, f (x, y, z) = 0.
Now we consider the extreme value of f on the boundary of R. Suppose that the extreme value
of f occurs at (x0 , y0 , z0 ). Note that the boundary of R consists of three pieces: g = 0, h = 0 and
g = h = 0.
1. g(x0 , y0 , z0 ) = 0: Since (∇g)(x0 , y0 , z0 ) ‰ 0, Lagrange Multiplier Theorem implies that there
exists λ P R such that
(y0 z0 , x0 z0 , x0 y0 ) = λ(2x0 , 2y0 , 2z0 ) .
Therefore, (x0 , y0 , z0 , λ) satisfies

y0 z0 = 2λx0 , (0.1a)
x0 z0 = 2λy0 , (0.1b)
x0 y0 = 2λz0 , (0.1c)
x20 + y02 + z02 = 4 . (0.1d)
 If one of x0 , y0 , z0 is zero, then f (x0 , y0 , z0 ) = 0; thus we assume that x0 y0 z0 ‰ 0. Then λ ‰ 0
and the product of (??a,b,c) shows that x0 y0 z0 = 8λ3 . Therefore,

4λ2 4λ2 4λ2

x0 = , y0 = , z0 = .
x0 y0 z0
which implies that (x0 , y0 , z0 ) is

(˘2λ, ˘2λ, 2λ) , (˘2λ, ¯2λ, ´2λ) , (˘2λ, 2λ, ˘2λ) ,

(˘2λ, ´2λ, ¯2λ) , (2λ, ˘2λ, ˘2λ) , (´2λ, ˘2λ, ¯2λ) .
1
In either cases, (??d) implies that 12λ2 = 4; thus λ = ˘ ? . Since z0 ě 1, we conclude that
3
( 2 2 2 ) ( 2 2 2 )
(x0 , y0 , z0 ) = ˘ ? , ˘ ? , ? or ˘ ? , ¯? , ? .
3 3 3 3 3 3
8
In this case, f (x0 , y0 , z0 ) = ˘ ? .
3 3
2. h(x0 , y0 , z0 ) = 0: Since (∇h)(x0 , y0 , z0 ) ‰ 0, Lagrange Multiplier Theorem implies that there
exists λ P R such that

(y0 z0 , x0 z0 , x0 y0 ) = λ(0, 0, 1) and z0 = 1 .

Therefore, (x0 , y0 , z0 ) = (0, 0, 1) which is impossible f (x0 , y0 , z0 ) = 0.

3. g(x0 , y0 , z0 ) = h(x0 , y0 , z0 ) = 0: Since

(∇g)(x0 , y0 , z0 ) ˆ (∇h)(x0 , y0 , z0 ) = (2x0 , 2y0 , 2z0 ) ˆ (1, 1, 1) = 2(y0 ´ z0 , z0 ´ x0 , x0 ´ y0 ) ,

(∇g)(x0 , y0 , z0 ) ˆ (∇h)(x0 , y0 , z0 ) = 0 if and only if x0 = y0 = z0 . Since h(x0 , y0 , z0 ) = 0 implies

that z0 = 1, and g(1, 1, 1) ‰ 0, we find that (∇g)(x0 , y0 , z0 ) ˆ (∇h)(x0 , y0 , z0 ) = 0. Therefore,
Lagrange Multiplier Theorem implies that there exist λµ P R such that

(y0 z0 , x0 z0 , x0 y0 ) = λ(2x0 , 2y0 , 2z0 ) + µ(0, 0, 1) .

Therefore, (x0 , y0 , z0 , λ, µ) satisfies

y0 z0 = 2λx0 , (0.2a)
x0 z0 = 2λy0 , (0.2b)
x0 y0 = 2λz0 + µ , (0.2c)
x20 + y02 + z02 = 4 , (0.2d)
z0 = 1 . (0.2e)

By (??a,b,e), we find that x0 = 2λy0 = 4λ2 x0 ; thus x0 = 0 or 4λ2 = 1.

(a) If x0 = 0, then f (x0 , y0 , z0 ) = 0.

 1
(b) If x0 ‰ 0, then λ = ˘ .
2
( 3 3 )
c c
1
(i) λ = : (??a,e) implies that y0 = x0 ; thus (??) implies that (x0 , y0 , z0 ) = , ,1 .
2 2 2
3
In this case, f (x0 , y0 , z0 ) = .
2
1 (
(ii) λ = ´ : (??a,e) implies that y0 = ´x0 ; thus (??) implies that (x0 , y0 , z0 ) = ˘
2
3 )
c c
3 3
,¯ , 1 . In this case, f (x0 , y0 , z0 ) = ´ .
2 2 2

Comparing the values of all possible extreme points (x0 , y0 , z0 ), we find that the maximum of f on
8 8
R is ? , and the minimum of f on R is ´ ? . ˝
3 3 3 3