1.
(a) State the defined meaning of a function f : A → R being differentiable
at a point a ∈ A.
(b) For f (x) = x1/3 arctan x2/3 , determine whether f is differentiable at
0.
Solutions:
(a) f being differentiable at a means that the limit
f (a + h) − f (a)
lim ,
h→0 h
exists.
(b) With f (x) = x1/3 arctan x2/3 we have
f (0 + h) − f (0) h1/3 arctan h2/3 arctan h2/3 − 0
lim = lim = lim
h→0 h h→0 h h→0 h2/3
1 2 −1/3
4/3 · 3 h
0 1
= = lim 1+h 2 −1/3 = lim = 1.
0 h→0 h h→0 1 + h4/3
3
Hence f is differentiable at 0 and f 0 (0) = 1.
1
2. Is the following statement true or false:
x 1
” 2
≤ , for all x ≥ 1.”
1+x 2
Justify your answer with a proof or a counterexample.
Solution: The statement is true. Here’s a proof:
Letting
x
f (x) = ,
1 + x2
we have
1
f (1) = ,
2
and
(1 + x2 ) · 1 − x · 2x 1 − x2
f 0 (x) = 2 2
= < 0, for x > 1.
(1 + x ) (1 + x2 )2
Hence f is (strictly) decreasing on [1, ∞), and in particular
1
f (x) ≤ f (1) = , for x ≥ 1.
2
Comment: The given inequality holds not just on [1, ∞) but indeed on all
of R, and there are a range of alternative ways to establish this.
2
3. (a) Consider f : A → R, A ⊂ R.
(i) What is the defined meaning of a point x0 ∈ A being a point of
local maximum for f .
(ii) What is the defined meaning of a point x0 ∈ A being a point of
global maximum for f .
(b) For each n = 1, 2, 3, . . ., consider the function fn : [0, 1] → R given
by
nx
fn (x) = .
1 + n2 x2
(i) For each x ∈ [0, 1], calculate the limit
lim fn (x),
n→∞
or explain why the limit does not exist.
(ii) Does each function fn have a point xn of global maximum? Ex-
plain your answer. If so, calculate the limit
lim fn (xn ),
n→∞
or explain why the limit does not exist.
Solutions:
(a) (i) x0 ∈ A being a point of local maximum for f means that there
exists some open interval (a, b) such that x0 ∈ (a, b) and
f (x0 ) ≥ f (x), for all x ∈ (a, b) ∩ A.
(ii) x0 being a point of global maximum for f means that
f (x0 ) ≥ f (x), for all x ∈ A.
(b) (i) For x = 0, we have
0
lim fn (x) = lim = 0,
n→∞ n→∞ 1+0
and for x ∈ (0, 1]:
nx 1 1 1
lim fn (x) = lim = lim · 1 = 0 · 2 = 0.
n→∞ n→∞ 1 + n2 x2 n→∞ n
n2 + x2 x
Hence
lim fn (x) = 0, for all x ∈ [0, 1].
n→∞
(ii) Since each fn is continuous on the closed bounded interval [0, 1],
the max/min theorem tells us that each fn has a point of global
maximum, which we may denote xn . Let’s find the global max-
imum value f (xn ):
We know that a global maximum value exists, and that xn must
be either a critical point of f or an endpoint of [0, 1]. Let’s check
3
the function values at these points. The largest value will be our
global maximum value. We have
(1 + n2 x2 ) · n − nx · 2n2 x n(1 − n2 x2 )
fn0 (x) = 2 2 2
= .
(1 + n x ) (1 + n2 x2 )2
The derivative is defined everywhere on [0, 1] and for x ∈ [0, 1],
n = 1, 2, . . ., we have
1
fn0 (x) = 0 ⇔ 1 − n2 x2 = 0 ⇔ x= .
n
Hence, x = 1/n is the only critical point of fn . Comparing the
function values at the critical point and endpoints (using the
result from Qn. 2), we see that
1 n 1
fn (0) = 0, fn (1/n) = , fn (1) = ≤ , for n = 1, 2, . . . .
2 1 + n2 2
Hence xn = 1/n is a point of global maximum for fn and
1 1
lim fn (xn ) = lim = .
n→∞ n→∞ 2 2
4
4. A solid object occupies a region in space given by
0 ≤ z ≤ 4 − x2 − y 2 .
Calculate the volume of this object.
Solution:
Note first that
4 − x2 − y 2 ≥ 0 ⇔ x2 + y 2 ≤ 4.
At a fixed radius r = x2 + y 2 from the origin in the xy-plane, the object
p
has a fixed height z = 4 − r2 . Considering a thin cylindrical shell centred
at the z-axis, with radius r, thickness dr, and height z = 4−r2 , its volume
is
dV = 2πr(4 − r2 ) dr,
and the total volume of the object is
2 2 2
r4
Z Z
16
V = dV = 2πr(4−r2 ) dr = 2π 2r2 − = 2π 8 − = 8π.
r=0 0 4 0 4
5
5. Solve the initial value problem in y = y(x):
4
y0 − y = −x3 , for all x > 0, y(1) = 2.
x
Solution:
The d.e. is a linear first order equation with integrating factor e−4 ln x =
x−4 . Multiplying both sides of the eqn. with the i.f. we get
x−4 y 0 − 4x−5 y = −x−1 ,
i.e.
d
x−4 y = −x−1 ,
dx
so Z
1
x−4 y = − dx = − ln x + C,
x
and we get
y = −x4 ln x + Cx4 .
The initial condition tells us that
2 = C,
so the solution to our i.v.p. is
y = −x4 ln x + 2x4 .
6
6. (a) Evaluate the limit
cos x − x−2 ln(1 + x2 )
lim .
x→0 (sin x)4
(b) Suppose f, g : (−1, 1) → R satisfy the following conditions.
(i) f and g are both continuous at 0.
(ii)
f (x) = axp + O(xp+ε ), g(x) = bxq + O(xq+ε ), as x → 0,
where p, q, ε > 0, p + q = 1.
Can we conclude that for h(x) = f (x)g(x) we have h0 (0) = ab?
Justify your answer with a proof or a counterexample.
Reminders:
∞
xn
for x ∈ (−1, 1],
X
ln(1 + x) = (−1)n−1 ,
n=1
n
∞
x2n+1
for x ∈ R,
X
sin x = (−1)n ,
n=0
(2n + 1)!
∞
x2n
for x ∈ R.
X
cos x = (−1)n ,
n=0
(2n)!
Solutions:
(a) For x 6= 0, |x| < 1, we have
cos(x) − x−2 ln(1 + x2 )
=
(sin x)4
x2 x4 (x2 )2 (x2 )3
1− 2 + 4! + O(x6 ) − x−2 x2 − 2 + 3 + O((x2 )4 )
=
(x + O(x3 ))4
x2 x4 x4 x6
1− 2 + 4! + O(x6 ) − x−2 x2 − 2 + 3 + O(x8 )
=
x4 (1 + O(x2 ))4
x2 x4 2
x4
1− 2 + 4! + O(x6 ) − 1 − x2 + 3 + O(x6 )
=
x4 (1 + O(x2 ))4
1 1 −7
− 3 x4 + O(x6 ) 2
4! + O(x ) −7 −7
4!
4 2 4
= 2 4
→ = , as x → 0.
x (1 + O(x )) (1 + O(x )) 4! 24
(b) The conclusion holds. Here’s a proof:
Note first that since f is continuous at 0, we have
f (0) = lim f (x) = lim [axp + O(xp+ε )] = 0,
x→0 x→0 | {z }
→0 since p,ε>0
and by a similar calculation, g(0) = 0.
7
Furthermore
f (x)g(x) = (axp + O(xp+ε ))(bxq + O(xq+ε )) =
abxp+q + axp O(xq+ε ) + bxq O(xp+ε ) + O(xp+ε )O(xq+ε ) =
abx + O(x1+ε ),
so
f (t)g(t) − f (0)g(0) abt + O(t1+ε )
h0 (0) = lim = lim = lim (ab+O(tε )) = ab.
t→0 t t→0 t t→0