MATH4007 Calculus 2 (M05-09) Final Examination Solution NTU 113-1

Date : 14th December, 2024

The total score for this exam is 100 points. This exam will contribute a maximum of 50% to your final grade
of MATH4007 Calculus 2.

Section A. Short Questions

A1 (6%) If f is a continuous function on R such that
x3
(x − 1) cos2 (πx) = ∫ f (t) dt,
1

find the value of f (8).

Solution:
Differentiate both side w.r.t x and use FTC

cos2 (πx) − 2(x − 1) sin(πx) cos(πx) = f (x3 ) ⋅ 3x2

Plug in x = 2, we obtain 1 = 12 ⋅ f (8) so f (8) = 1/12.

Scoring:

ˆ Correctly perform diffentiation and apply FTC: [4%] (2% for correctly differentiate LHS and 2% for
correctly apply FTC.)
ˆ Correct evaluation at x = 2: [2%] (1% if the student know he should evaluate at x = 2 but answer is wrong.)

1
A2 (6%) Find ∫ √ dx.
4x2 + 4x + 5

Solution:
1
∫ √ dx , 2x + 1 = 2 tan θ , θ ∈ (−π/2, π/2)
4x2 + 4x + 5
1
=∫ sec2 θ dθ
2 sec θ
1
= ln ∣ sec θ + tan θ∣ + C
2
√
1 4x2 + 4x + 5 2x + 1
= ln ( + )+C
2 2 2

Grading:
There could be lots of possible answers for this problem. We’ll simplify grading as -2% per mistake and -4% for
each incorrect use of integration technique. (or more simply, -2 for minor and -4 for major mistakes)
We do not give points for correct final answer unless they clearly showed the derivative of the answer is the
original problem, which is work to justify their answer.
If students try several different methods, grade the one with the least amount of mistakes. -2% if they did not
label which one is their final version.
If the student did not reach an answer for the problem, then we award them +2% for each correct technique
applied to the integral that leads them closer to the answer.

Page 1 of 12
 e
A3 (6%) Find ∫ (x ln x)2 dx.
1

Solution:
e
2
∫ (x ln x) dx
1
e
=∫ x2 (ln x)2 dx
1
e
x3 1 e
=[ (ln x)2 ] − ∫ 2x2 ln x dx
3 1
3 1
e
e3 2 x3 1 e
= − [[ ln x] − ∫ x2 dx]
3 3 3 1
3 1
e3 2e3 2(e3 − 1) 5e3 − 2
= − + =
3 9 27 27

Or e
2
∫ (x ln x) dx , u = ln x
1
1
=∫ u2 e3u du
0
1
u2 e3u 2ue3u 2e3u 5e3 − 2
=[ − + ] =
3 9 27 0 27

Grading:
There could be lots of possible answers for this problem. We’ll simplify grading as -2% per mistake and -4% for
each incorrect use of integration technique. (or more simply, -2 for minor and -4 for major mistakes)
We do not give points for correct final answer unless they clearly showed the derivative of the answer is the
original problem, which is work to justify their answer.
If students try several different methods, grade the one with the least amount of mistakes. -2% if they did not
label which one is their final version.
If the student did not reach an answer for the problem, then we award them +2% for each correct technique
applied to the integral that leads them closer to the answer.

ex
A4 (8%) Find ∫ dx.
(ex + 1)2 (e2x + 1)

Solution:
ex
∫ dx , u = ex
(ex + 1)2 ⋅ (e2x + 1)
1
=∫ du
(u + 1)2 ⋅ (u2 + 1)
A B Cu + D
=∫ + + 2 du
u + 1 (u + 1)2 u +1
System of equations: A + C = 0, A + B + 2C + D = 0, A + C + 2D = 0, A + B + D = 1. Set u = −1 to get 2B = 1.
Solution is A = 12 , B = 21 , C = − 12 , D = 0.

1 1 1 u
= ∫ + − du
2 u + 1 (u + 1)2 u2 + 1
1 1 1
= ln(ex + 1) − x
− ln(e2x + 1) + C
2 2(e + 1) 4

Grading:

Page 2 of 12
 There could be lots of possible answers for this problem. We’ll simplify grading as -2% per mistake and -4% for
each incorrect use of integration technique. (or more simply, -2 for minor and -4 for major mistakes)
We do not give points for correct final answer unless they clearly showed the derivative of the answer is the
original problem, which is work to justify their answer.
If students try several different methods, grade the one with the least amount of mistakes. -2% if they did not
label which one is their final version.
If the student did not reach an answer for the problem, then we award them +2% for each correct technique
applied to the integral that leads them closer to the answer.

1 1 1
A5 Let R be the region bounded by the curves y = x2 − ln x, y = x2 , and x = e. (See below)
2 4 2

b
(a) (4%) The integral ∫ f (x)dx expresses the volume of the solid obtained by rotating R about the x-axis. Find
a
a, b and f (x). (You don’t need to calculate the integral.)
d
(b) (4%) The integral ∫ g(x)dx expresses the volume of the solid obtained by rotating R about x = −2. Find c, d
c
and g(x). (You don’t need to calculate the integral.)
1 1
(c) (6%) Calculate the arclength of C, the segment of the curve y = x2 − ln x that forms part of the boundary of
2 4
R.

Solution:

(a) First, we determine the x-limits of the region. The two curves intersect where:
1 2 1 2 1 1
x = x − ln x Ô⇒ − ln x = 0 Ô⇒ ln x = 0 Ô⇒ x = 1.
2 2 4 4
Thus, the region extends from x = 1 to x = e. Therefore, a = 1 and b = e.
When rotating about the x-axis, we use the washer method. The outer radius is given by youter = 12 x2 , and
the inner radius is yinner = 21 x2 − 14 ln x.
The volume element is:
⎡ 2 2 2⎤
2 2
⎢ x
⎢ x2 ln x ⎥⎥
π [(outer radius) − (inner radius) ] = π ⎢( ) − ( − ) ⎥.
⎢ 2 2 4 ⎥
⎣ ⎦
2 ln x ln x
= π (x − )
4 4

Page 3 of 12
 Hence:
ln x ln x
f (x) = π (x2 − ) .
4 4
Answer for (a):
ln x ln x
a = 1, b = e, f (x) = π (x2 − ) .
4 4

Partial Credit (4% total):

ˆ Correctly identifying a = 1 and b = e: 2%
ˆ Correct integrand form f (x) (simplification is not necessary): 2%
(b) Rotating around the vertical line x = −2, we use the cylindrical shell method. The region in x is still from
x = 1 to x = e, so c = 1 and d = e.
A representative vertical strip at x has:

radius = x − (−2) = x + 2,
1 1 1 1
height = ytop − ybottom = x2 − ( x2 − ln x) = ln x.
2 2 4 4
Using shells, the volume element is:
1 π
2π(radius)(height) dx = 2π(x + 2) ( ln x) dx = (x + 2) ln x dx.
4 2

Thus:
π
g(x) = (x + 2) ln x.
2
Answer for (b):
π
c = 1, d = e, g(x) = (x + 2) ln x.
2

Partial Credit (4% total):

ˆ Correctly identifying c = 1 and d = e: 2%
ˆ Correct integrand form g(x): 2%
(c) First, find the derivative:
1 1
y(x) = x2 − ln x.
2 4
1 4x2 − 1
y ′ (x) = x − = .
4x 4x
The arclength from x = 1 to x = e is:
e√
L=∫ 1 + (y ′ (x))2 dx.
1

Compute (y ′ (x))2 :
2
′ 24x2 − 1 (4x2 − 1)2
(y (x)) = ( ) = .
4x 16x2
Thus:
(4x2 − 1)2
1 + (y ′ (x))2 = 1 + .
16x2
Expand (4x2 − 1)2 = 16x4 − 8x2 + 1, so:

16x2 + (16x4 − 8x2 + 1) 16x4 + 8x2 + 1

1 + (y ′ (x))2 = = .
16x2 16x2

Notice that 16x4 + 8x2 + 1 = (4x2 + 1)2 . Therefore:

√ 4x2 + 1
1 + (y ′ (x))2 = .
4x

Page 4 of 12
 Hence:
e 4x2 + 1 e 1
L=∫ dx = ∫ (x + ) dx.
1 4x 1 4x
Integrate:
e x2 e e2 1
∫ x dx =
∣ = − ,
1 2 1 2 2
e 1 1 e 1 1
∫ dx = ln x∣ = (ln e − ln 1) = .
1 4x 4 1 4 4
Adding these:
e2 1 1 e2 1
L=( − )+ = − .
2 2 4 2 4

Answer for (c):

e2 1
L= − .
2 4

Partial Credit (6% total):

ˆ Correct derivative y ′ (x): 2%
ˆ Correct integrand in arclength formula: 2%
ˆ Correct final evaluation: 2%

B1 The principal of the Munch High School wants to study the economic profile of the families of students. Let X (in
ten thousand dollars) be the family annual income of a randomly selected student. It is known that the (cumulative)
probability distribution function of X is given by
⎧
⎪a − be−0.1x
⎪ if x ≥ 0
F (x) = P(X ≤ x) = ⎨ .
⎪
⎪0 if x < 0
⎩
You are given that F (0) = 0.

(a) (3%) Find the values of a and b.

Since lim F (x) = 1, we obtain a = 1.
x→∞
By using the given condition F (0) = 0, we have a − b = 0.
Therefore, we deduce that a = b = 1.
Grading scheme.
ˆ (1M) Mentioning lim F (x) = 1.
x→∞

ˆ (1M) Correct value of a

ˆ (1M) Correct value of b

Remark. Partial credit : 1M for those who only manage to obtain a = b by using the given
condition. In both (a),(b), we may deduct marks from students with correct answers but incorrect
and/or incomplete reasoning.
(b) (3%) Find a probability density function f (x) of the random variable X.
⎧
⎪0.1e−0.1x if x ≥ 0;
⎪
f (x) = ⎨ .
⎪
⎪0 if x < 0
⎩
Grading scheme.
ˆ (1M) The case for x < 0;

ˆ (2M) The case for x ≥ 0.

(c) (8%) The principal is preparing for the budget for the next academic year. The principal classifies the students
into three categories and subsidizes students of different categories by a different amount of academic scholarship.

Page 5 of 12
 Category Family Annual Income x Subsidy S(x)
A 0≤x≤1 $ 500
B 1≤x≤4 $ (650 − 150x)
C x≥4 $ 50
∞
Find the expected subsidy ∫ S(x)f (x) dx of a randomly selected student from school.
0

∞ 1 4 ∞
∫ S(x)f (x)dx = ∫ 500 ⋅ 0.1e−0.1x dx + ∫ (650 − 150x)0.1e−0.1x dx + ∫ 50 ⋅ 0.1e−0.1x dx .
0 0 1 4
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
I1 I2 I3

Now,

I1 = 500 ⋅ F (1) = 500(1 − e−0.1 ) = 500 − 500e−0.1 ;

4
I2 = 650 ⋅ (F (4) − F (1)) − 15 ∫ xe−0.1x dx = 650(e−0.1 − e−0.4 ) − 15[−10xe−0.1x − 100e−0.1x ]x=4
x=1
1
= 650(e−0.1 − e−0.4 ) − 15(110e−0.1 − 140e−0.4 )
= 1450e−0.4 − 1000e−0.1 ;
I3 = 50(1 − F (4)) = 50e−0.4 .

Adding up all these, we obtain I1 + I2 + I3 = 500 − 1500e−0.1 + 1500e−0.4 .

Grading scheme.

ˆ (1M) Setting up the correct integrals (as sum of three integrals)

ˆ (1.5+1.5M) Correct values of I1 and I3 (Partial credit : 1M for the indefinite integral of e−0.1x )

ˆ (3M) Correct value of I2 (Partial credit : 2M for the indefinite integral of xe−0.1x )

ˆ (1M) Correct answer

Page 6 of 12
B2 In a biochemical research, understanding the rate of substrate degradation is crucial for drug metabolism studies.
Consider a scenario where a drug (substrate s) is broken down by an enzyme in the body. In addition to enzyme-
catalyzed degradation, the substrate is also lost at a constant rate due to leakage. The dynamics of the system can
be described by the following differential equation:
dCs
= −kCs − l, where
dt
● Cs = Cs (t) is the concentration of the substrate (in millimolar, mM) at time t (in minutes),
● k = 0.1 min−1 is the rate constant for enzyme-mediated degradation,
● l = 1 mM/min represents the constant substrate loss due to leakage.
Initially, the substrate concentration is Cs (0) = 50 mM.
(a) (6%) Calculate the concentration of the drug remaining in the body after 10 minutes.
(b) (4%) Determine the time required for the drug concentration to decrease to 5 mM.
(c) (5%) A pharmacologist claims that if the rate constant k is doubled from 0.1/min to 0.2/min (while keeping
l and Cs (0) to be the same as above), the time required for the drug concentration to drop to 5 mM will be
halved. Is this claim correct? Justify your reasoning mathematically.

Solution:

(a) Separable equation method,

dCs
∫ = − ∫ dt
0.1Cs + 1
ln(0.1Cs (t) + 1)
integrate both side we obtain : = −t + A, where A is arbitrary constant.
0.1
Plug in initial condition Cs (0) = 50 to solve for A = 10 ln 6.
Simplify the expression Cs (t) = 60e−0.1t − 10. Hence Cs (10) = 60e−1 − 10.

Alternatively, student can also solve it using Integrating factor method:

dCs d 0.1t
+ 0.1Cs = −1 Ô⇒ (e Cs ) = −e0.1t
dt dt
So Cs (t) = −e−0.1t ∫ e0.1t dt = −10 + Ae−0.1t where A is arbitrary constant. Plug in initial value to solve for
A then Cs (t) = 60e−0.1t − 10.
(b) Solve Cs (t) = 5. Using the solution we obtain in part (a). Then

5 = 60e−0.1t − 10 Ô⇒ t = 10 ln 4

Any equivalent form of answer is accepted i.e. t = 20 ln 2, − ln(15/60)/0.1, ln(4)/0.1, 13.86.

(c) We need to solve the differential equation using k = 0.2. The equation to be solved is
dCs
= −0.2Cs − 1, Cs (0) = 50
dt
Solve it using either separable or integrating factor method: Cs (t) = 55e−0.2t − 5. Now we find t for which
Cs (t) = 50 × 10% = 5. The corresponding value of t is : t = 5 ln(11/2).
Comparing this answer with the one in part (b) , we find out that 1/2 ⋅ 10 ln 4 ≠ 5 ln(11/2). Hence the
pharmacologist claim is FALSE.

Scoring: Part (a)

ˆ Correctly solve the differential equation. Allow to not simplify the answer :[4%]
ˆ Correctly use the initial condition to determine the constant: [1%]
ˆ Correctly evaluate Cs (10). Allow to not simplify the answer: [1%]

Part (b) Conceptually understand the question wants to solve for Cs (t) = 5 (even if student solve (a) incorrectly):
[2%]. Correctly solve for t : [2%]
Part(c) No points will be giving if it is pure guessing for any correct conclusion (claim is FALSE).

Page 7 of 12
 ˆ set up the correct IVP with k = 0.2:[1%]

ˆ corectly solve the IVP. Allow to not simplify the answer. [2%]

ˆ correctly solve for t , in Cs (t) = 5 [1%]

ˆ correct conclusion (FALSE) after comparing the t obtained with the one in part (b): [1%] .

Alternatively, student can also solve both the initial value problem of (a),(c) using general constant k then plug
in value for different k in different question. If they can solve the initial value problem in part (a)
correctly, then it is allowed in part (c) to directly write down the answer for the IVP.

Page 8 of 12
B3 Let P = P (t) represent the population of a species at time t (days). Consider the differential equation
dP P
= k(t) ⋅ P ⋅ (1 − ),
dt M (t)
where k(t) and the carrying capacity M (t) ≠ 0 are differentiable functions of t representing seasonal variation.
1 1
(a) Suppose M (t) = M is a constant and k(t) = + . In addition, P (1) = p0 where 0 < p0 < M .
2 t
(i) (6%) Solve for P = P (t), with t > 0, in this case.
(ii) (2%) Find lim P (t) in terms of M .
t→∞
(b) Now we explore the case when M = M (t) is not a constant. Assume P (t) ≠ 0.
1
(i) (3%) Verify that the substitution z = transforms the given equation into a first order linear equation
P
dz k(t)
+ k(t)z = .
dt M (t)
(ii) (8%) Suppose M (t) = A − e−kt and k(t) = k where both A > 1, k > 0 are constants. Solve for z = z(t), with
t ≥ 0, in this case.
(iii) (2%) In the setting of (b)(ii), find lim P (t) in terms of A.
t→∞

Solution:

(a) (i) It is a separable equation.

dP 1 1 P M 1 1
= ( + ) ⋅ P ⋅ (1 − ) Ô⇒ ∫ dP = ∫ + dt
dt 2 t M P (M − P ) 2 t

1 1 t P t
Ô⇒ ∫ + dP = + ln t + C Ô⇒ ln ∣ ∣ = + ln t + C
P M −P 2 M −P 2
P
Because 0 < P (1) = p0 < M , > 0 for t near 1. Hence for t near 1,
M −P
P t
= A ⋅ te 2 where A > 0 is some constant .
M −P
Hence
M
P (t) = t .
1 + A−1 t−1 e− 2
√ M − p0
Since P (1) = p0 , we can solve A−1 = e . Therefore,
p0
M
P (t) = M −p0 −1 − t−1
.
1+ p0
t e 2

Note that for t > 0, the solution P (t) exists and P (t) < M .

Grading Criteria:
M 1 1
1 pt separate P and t into two sides of the equation ∫ dP = ∫ + dt.
P (M − P ) 2 t
1 t
1 pt for ∫ dt = + ln t + C.
t 2
M P
2 pts for ∫ dP = ln ∣ ∣ + C.
P (M − P ) M −P
M
1 pt for P (t) = t , where A > 0 is a constant.
1 + A t−1 e− 2
−1
1 pt for the constant A.
t−1 M
(ii) Because lim t−1 e− 2 = 0, we obtain that lim P (t) = = M.
t→∞ t→∞ 1+0
Grading Criteria:
2 pts for lim P (t) = M .
t→∞

Page 9 of 12
 1 P ′ (t)
(b) (i) Let z(t) = . Then z ′ (t) = − . Dividing the differential equation by (P (t))2 , we obtain
P (t) (P (t))2

P ′ (t) 1 1 k(t)
2
= k(t) ( − ) Ô⇒ −z ′ (t) = k(t)z(t) − .
(P (t)) P M (t) M (t)
Hence
dz k(t)
+ k(t)z(t) = .
dt M (t)
Grading Criteria:
P ′ (t)
1 pt for z ′ (t) = − .
(P (t))2
dz k(t)
2 pts for rewriting the differential equation as + k(t)z = .
dt M (t)
(ii) Suppose M (t) = A − e−kt and k(t) = k where both A > 1, k > 0 are constants. Then
dz k
+ kz(t) = .
dt A − e−kt
Choose the integrating factor I(t) = ekt . Multiply the above differential equation by I(t) and we have

dz kekt
ekt + kekt z(t) = (ekt z(t))′ = .
dt A − e−kt
integrating the both sides, we obtain
kekt u=ekt 1 u
ekt z(t) = ∫ dt = ∫ 1
du = ∫ du
A−e −kt A− u
Au − 1
1 1 u 1
= ∫ 1+ du = + 2 ln ∣Au − 1∣ + C
A Au − 1 A A
1 1
= ekt + 2 ln ∣Aekt − 1∣ + C
A A
where C is a constant. Hence
1 1
z(t) = + 2 e−kt ln ∣Aekt − 1∣ + Ce−kt ,
A A
where C is a constant.

Grading Criteria:
1 pt for the integrating factor I(t) = ekt .
1 pt for multiplying the differential equation with I(t).
kekt 1 1
5 pts for ∫ dt = ekt + 2 ln ∣Aekt − 1∣ + C. Partial credits as follows:
A − e−kt A A
1 pt for choosing u = ekt .
kekt 1
1 pt for ∫ dt = ∫ du.
A − e−kt A − u1
u 1 1
1 pt for = (1 + )
Au − 1 A Au − 1
1 pt for ∫ 1 du = u + C.
1 1
1 pt for ∫ du = ln ∣Au − 1∣ + C.
Au − 1 A
1 1
1 pt for z(t) = + 2 e−kt ln ∣Aekt − 1∣ + Ce−kt ,
A A
Akekt
kt ln ∣Aekt − 1∣ ∞ kt A 1
−kt
(iii) Note that lim e ln ∣Ae −1∣ = lim kt
= lim Ae kt−1 = lim
∞
kt
= 0. Hence lim z(t) = .
t→∞ t→∞ e t→∞ ke t→∞ Ae − 1 t→∞ A
1
Thus lim P (t) = lim = A.
t→∞ t→∞ z(t)

Grading Criteria:
1 pt for lim e−kt ln ∣Aekt − 1∣ = 0.
t→∞
1 pt for lim P (t) = A.
t→∞

Page 10 of 12
B4 Suppose a is a constant.
(a) (6%) Determine the range of a such that the improper integral
∞ 1
Ia = ∫ dx
1 xa + x2−a
converges based on comparison test with suitable p-integrals.
1
(b) (4%) By considering the substitution u = on Ia , determine the range of a such that the improper integral
x
∞ 1
Ja = ∫ dx converges.
0 x + x2−a
a

Solution:
(a) We consider the behavior of the integrand
1
fa (x) =
xa + x2−a
as x → ∞.
First, identify the dominant term in the denominator for large x. Compare the exponents a and (2 − a):

ˆ If a > 1, then a > 2 − a (since a > 1 Ô⇒ 2 − a < 1). Thus, for large x, xa dominates and
1
0 < fa (x) < .
xa
∞ 1
A p-integral ∫1 xp
dx converges if and only if p > 1. Here, we require a > 1 for convergence as x → ∞.
ˆ If a < 1, then 2 − a > 1, so x2−a dominates for large x. Thus
1
0 < fa (x) < .
x2−a
∞ 1
The integral ∫1 x2−a dx converges if 2 − a > 1 Ô⇒ a < 1. This matches our assumption a < 1, so for a < 1,
the integral also converges at infinity.
ˆ If a = 1, the integrand becomes
1 1 1
f1 (x) = = = ,
x1 +x 2−1 x+x 1 2x
∞ 1
and ∫1 2x
dx diverges.

No issues arise at the lower limit x = 1, since the integrand is finite there.
Combining these findings, the integral Ia converges if and only if a < 1 or a > 1. It diverges at a = 1.
Answer for (a): Ia converges for a < 1 or a > 1, and diverges for a = 1.

(b) Consider
∞ 1
Ja = ∫ dx.
0 xa + x2−a
Split the integral at x = 1:
1 1 ∞ 1
Ja = ∫ dx + ∫ dx.
0 xa +x 2−a 1 xa + x2−a
∞ 1
The second integral, ∫1 xa +x2−a
dx, is exactly Ia , whose convergence condition we already know.
1 1
Now, focus on the first integral ∫0 xa +x2−a
dx. Make the substitution:

1 1 1
u= Ô⇒ x = , dx = − du.
x u u2

When x = 0+ , we have u → ∞, and when x = 1, we have u = 1. Thus:

1 1 1 1 1
∫ dx = ∫ a 2−a
(− 2 ) du.
0 xa + x2−a ∞ (1) +(1) u
u u

Page 11 of 12
 Simplify the denominator:
1 a 1 2−a
( ) = u−a , ( ) = ua−2 .
u u
So the integrand becomes:
−1
.
u2 (u−a + ua−2 )

Change the limits accordingly and flip them to remove the negative sign:
1 1 ∞ 1
∫ dx = ∫ du.
0 xa +x 2−a 1 u + u2−a
a

Notice that this integral in u is of the same form as Ia , but just using a different variable of integration. Therefore,
1
the convergence of ∫0 portion is exactly the same condition as Ia .
We now have:
1 1 ∞ 1
Ja = ∫ dx + ∫ dx = 2Ia .
0 xa +x 2−a 1 x + x2−a
a

Hence the convergence condition for Ja is the same as the one for Ia :

Ja converges if and only if a < 1 or a > 1.

It diverges at a = 1.

Answer for (b):

Ja converges if and only if a < 1 or a > 1, and diverges for a = 1.

Partial Credit:

(a) (6%) – Correct analysis of behavior as x → ∞: 3%

– Correct conclusion on the range of a for convergence: 3%
(b) (4%) – Correct splitting at x = 1 and appropriate substitution u = 1/x: 2%
– Correct reduction of the (0, 1) integral to the known (1, ∞) case, and correct final conclusion: 2%

Page 12 of 12