(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

Final Examination (Take Home) Solution

MATH 1023: Honors Calculus I

Fall 2019, HKUST
February 17th, 2020, 16:30 - 19:30
• Due to quite a lot of students experienced technical difficulties on Canvas, we decide to extend
the late penalties by 15 minutes. If you are late for more than 15 minutes, please accept a
small penalty.
• We found Q1(a)(iv) to be a bit ambiguous: “not continuous” may refer to either f (x) or
f (1023) (x). We decide to accept answers to both approach.

Problem 1.
(a) (i) Does not exists. By Extreme Value Theorem, f (x) has global max and min on closed
interval, hence the range cannot be an open interval.
(ii) e.g. f (x) = sin2 (2⇡x)
P n2
(iii) f (x) = (x 1)xex 1 (f (x) = 1 n=1 n! (x 1)n also accepted.)
⇢ 2046
x sin x1 x 6= 0
(iv) e.g. f (x) = .
0 x=0
Edit: if f (1023) (x) exists, then f (x) is contiuous. We now accept the answer “not exists”
(with explanation) also since the question may be ambiguous.
( 2
( 1)N ( a + ax + ax2 + o(x2 )) x < 0
(b) f (x) = b 2 x2 since x3 D(xN ) = 0 + o(x2 ).
1 + bx 2 8 + o(x 2) x 0

Comparing both sides, we have a = ( 1)N +1 , b = 2a( 1)N = 2.

x 2
(c) f 0 (x) = xlog 2
1
log x2
p 1
log 2
Hence critical points are: x = 1, 2, 2 with values 1, 2 (< 1), 1. Hence
4

p 1
log 2
Global max: x = 1, 2 with value 1. Global min: x = 2 with value 2 4
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

(d) Rewrite
1 n 1
x2n = Pn (x)(x )(x ) + a n x + bn
n n
1
Putting x = n we have ( n1 )2n = n1 an + bn
1 1 2n 1
Putting x = 1 n we have (1 n) = (1 n )an + bn
Subtracting,
1 2n 1 2
(1 ) ( )2n = (1 )an
n n n
hence
1 2n
(1 n) ( n1 )2n
an = 2
1 n
and
e 2 0 2
lim an = =e
n!1 1

Then
1 1
lim bn = lim ( )2n an = 0 0e 2
=0
n!1 n!1 n n

(e) Rearranging, we have

2
a ab
 a2
b b
() log a log b  b2 log a a2 log b
log a log b
() 2

1 a 1 b2
log x
Hence it suffices to show that the function f (x) = 1 x2
is increasing on (0, 1).

1 x2 + 2x2 log x
f 0 (x) =
x(x2 1)2

It suffices to show g(x) = 1 x2 + 2x2 log x 0 on (0, 1].

Since g(1) = 0 and g 0 (x) = 4x log x < 0, we conclude that g(x) > 0.

2
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

(f) Assume limx!0+ f 0 (x) converges to L. Then there exists c such that |f 0 (x)| < |L| + 1 = M is
bounded for x 2 (0, c).
For x 2 (0, c), by Mean Value Theorem, there exists ⇠ 2 (x, c) ⇢ (0, c) such that

f (x) f (c)
| | = |f 0 (⇠)| < M
x c
Hence for x 2 (0, c), we have

|f (x)| < |f (x) f (c)| + |f (c)| < M |x c| + |f (c)| < M c + |f (c)|

is bounded for all x 2 (0, c), hence limx!0+ f (x) cannot diverge to infinity.

Alternatively: proving directly:

If limx!0+ f (x) diverges, then the sequence xn = f ( n1 ) diverges. Hence it is not Cauchy:
There exists ✏0 > 0 such that for any N , |xn xm | ✏0 for some n, m > N . Also | n1 1
m|  1
N.

By Mean Value Theorem, there exists cN 2 ( n1 , m

1
) ⇢ (0, N1 ) such that

f ( n1 ) 1
f(m )
|f 0 (cN )| = | 1 1 | N ✏0
n m

Hence f 0 (cN ) is a divergent sequence, so limx!0+ f 0 (x) diverges.

3
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

Problem 2.
(a) By L’Hopital’s Rule:

sinN x 1 cosN x 1
= lim
x!0 xN cosN x x2
N cosN 1 x sin x
= 0 1 · 1 · lim
0 x!0 2 x
N
=
2
sin x
since limx!0 x = 1.

Alternatively, by Taylor’s Expansion,

x3
sin x x + o(x3 ) x3 x2 x3
tan x = = 6
x2
= (x )(1 + ) + o(x3 ) = x + + o(x3 )
cos x 1 + o(x3 ) 6 2 3
2

Hence
x3 x3
sinN x tanN x = (x + o(x4 ))N (x + + o(x3 ))N
6 3
x2 x2
= xN (1 + o(x3 ))N (1 + + o(x2 ))N )
6 3
x2 x2
= xN (1 N (1 + N ) + o(x2 ))
6 3
N N +2 N +2
= x + o(x )
2
N
Hence the limit is 2.

(b) Rewrite with n = x1 , the limit equals

r r r r
1 1 1 1 1
= lim 3 ( +1 + +2 + 3)
x!0+
x2 x x x x
1 p p p
= lim 2 ( 1 + x 1 + 1 + 2x 1 + 3x)
x!0+ x
1 x x2 x2 3x 9x2
= lim 2 ((1 + ) 1 + (1 + x ) (1 + ) + o(x2 ))
x!0+ x 2 8 2 2 8
1 x2 x2 9x2
= lim 2 ( + ) + o(x2 ))
x!0+ x 8 2 8
1 1
= ( 1 4 + 9) =
8 2

4
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

(c) We can rewrite it as derivative. Let f (x) = ax bx . Then f (0) = 0 and

2 2 2 2
!2
ax bx ax bx 1
lim x = lim ax bx
x!0 (a bx )2 x!0 x2 x
!2
ah bh 1
= lim ax b x
h!0+ h
limx!0 x
1 1
= f 0 (0) 0 2
= 0
(f (0)) f (0)
1 1
Since f 0 (x) = ax log a bx log b, f 0 (0) = log a log b 6= 0. So the answer is = .
log a log b log ab

Alternatively, by L’Hôpital’s Rule:

2 2
2x ax log a bx log b
= 0 lim
0 2(ax
x!0 bx ) ax log a bx log b
x
= lim x
x!0 a bx
1
= 0 lim
0ax log a
x!0 bx log b
1
=
log a log b

5
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

Problem 3.
The results of this problem is known as Darboux Theorem, which states that f 0 (x) always satisfy
the Intermediate Value Property. This problem is adapted from the proof in wikipedia.
(a) [2] First, ↵(t) and (t) are continuous.
At the turning point t = m, we have limt!m+ ↵(t) = limt!m+ 2t b = 2m b = a = ↵(m)
and similarly limt!m+ (t) = b = 2m a = (m).
[2] Next, the denominator of g(t) is non-zero for t 2 (a, b):
If a < t  m, ↵(t) = a < 2t a = (t). If m  t < b, ↵(t) = 2t b < b = (t).
Hence ↵(t) < (t).
Alternatively, plot a graph to explain.

[4] Since g(t) is a rational function of the composition of continuous functions with nonzero
denominator, it is continuous on (a, b).
[4] Check the endpoints. We have
f (2t a) f (a) f (a + h) f (a)
lim g(t) = lim = lim = f 0 (a)
t!a+ t!a+ 2t 2a h!0+ h
by linear change of variable h = 2t 2a. Hence g(t) is continuous at t = a.
Similarly,
f (b) f (2t b) f (b + h) f (b)
lim g(t) = lim = lim = f 0 (b)
t!b t!b 2b 2t h!0 h
by linear change of variable h = 2t 2b. Hence g(t) is continuous at t = b.
(b) Since g(t) is continuous, by Intermediate Value Theorem, there exists d 2 (a, b) such that
g(d) =
i.e.
f ( (d)) f (↵(d))
=
(d) ↵(d)
Since a  ↵(d) < (d)  b (by the graph above), by Mean Value Theorem, there exists
c 2 (↵(d), (d)) ⇢ (a, b) such that
f 0 (c) =

6
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

Problem 4.
(a) Plotting the graph g(x) = x3 3x with g 0 (x) = 3(x 1)(x + 1) and g 00 (x) = 6x.

The local max is x = 1 with value g( 1) = 2, and local min is x = 1 with value g(1) = 2.
8
< one root p < 2 or p > 2
Hence g(x) = p has two roots p = 2 or p = 2 .
:
three roots 2<p<2
(b) If p < 2 or p > 2, by definition f (p) = ↵2 where ↵ is the only root.
When p = 2, g(x) = 2 has two roots x = 1, x = 2.
When p = 2, g(x) = 2 has two roots x = 2, x = 1
Hence from the graph above, we have |↵| > 2 and f (p) > 4.
Otherwise, let ↵   be the three roots (including repeated roots) of g(x) = p. Then

(x ↵)(x )(x ) = x3 3x p

implies
↵+ + = 0, ↵ +↵ + = 3
Hence
2
f (p) = ↵ = ↵ 3= (↵ + ) 3= 3

Hence f (p) attain its minimum when = 0, i.e. when p = 0 with value f (0) = 3.
(c) f (p) is an even function since g(x) is an odd function:
For p > 2, if g(↵) = p, g( ↵) = p, hence f (p) = ↵2 = f ( p) for p > 2.
For 2  p  2, if ↵   are the roots of g(x) = p, then   ↵ are the roots
of g(x) = p, hence f (p) = ↵ = ( )( ↵) = f ( p).
(d) Since f (p) is even, to plot the graph we only need to study the right half p 0.
Discontinuity: f (2) = 2 but limp!2+ f (p) = lim↵!2+ ↵2 = 4.

7
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

Slope: For 0  p  2, we have 3 3 = p. Hence by implicit di↵erentiation,

2d d d 1
3 3 = 1 =) =
dp dp dp 3 2 3
d d 2
f 0 (p) = ( 2
3) = 2 =
dp dp 3 2 3
Since 1   0, f 0 (p) > 0, the graph is increasing with infinite slope at p = 2 where
= 1.
2↵
For p > 2, f 0 (p) = 3↵2 3
> 0 is increasing.
Hence f (0) is the only local minimum and there are no local maximums.
Infinity and Asymptotes: As p ! +1, ↵ ! +1, f (p) = ↵2 ! +1.
Since limp!+1 f (p) = +1 but limp!+1 f 0 (p) = 0, it has no asymptotes.
Convexity: For 0 < p < 2,

(3 2 3)2 2 (6 ) d 2 2 +1
f 00 (p) = 2
= >0
(3 3)2 dp 9( 2 1)3

since 1  0, hence it is convex.

2 ↵2 +1
For p > 2, f 00 (p) = 9 (↵2 1)2 < 0 hence it is concave.

8
(MATH1023)[2019](f)final~=zd5muen^_49360.pdf downloaded by yympong from http://petergao.net/ustpastpaper/down.php?course=MATH1023&id=3 at 2021-09-22 12:48:52. Academic use within HKUST only.

Point Distribution:
– Graph: 7 points:
⇤ even function symmetry
⇤ global min at (0, 3)
⇤ discontinuity at (±2, 2) and (±2, 4)
⇤ Tends to infinity
⇤ Convex on |p| < 2
⇤ Concave on |p| > 2
⇤ infinite slope at (±2, 2)
– Other:
⇤ 2 x Show increasing by f 0 (p)
⇤ 2 x Show convexity by f 00 (p)
⇤ 1 x Show no asymptotes.