Assignment 1

Abhinav Pradeep
Tutorial group 5
August 11, 2023

1 Question 1.
sin(xy)
f (x, y) =
x+y
sin(xy)
g(x, y) =
xy

1.1 a
Code to plot f (x, y):

syms x y
f(x,y)= sin(x*y)/(x+y);
ezsurf(f)

Code to plot g(x, y):

syms x y
g(x,y)= sin(x*y)/(x*y);
ezsurf(g)

1
From produced graphs, it can be inferred that f (x, y) does not have a limit at (0, 0) whereas g(x, y)
does. The discontinuity of f (x, y) appears to be along the curve defined by x = −y. This is sensible
as when f (x, y) approaches the curve x = −y, the denominator x + y approaches 0, making f (x, y)
discontinuous everywhere along the curve.

1.2 b
Based on the produced graph, the lim(x,y)→(0,0) f (x, y) appears to not exist. As the function is entirely
discontinuous along x = −y, there exists no ’paths’ along that line that would terminate in any
non-undefined value.
In other words, continuity (as defined by the existence of a limit at the point) requires that
p
∀ϵ > 0, ∃δ > 0 such that if 0 < |(x, y) − (0, 0)| ⇔ x2 + y 2 < δ

|f (x, y) − 1| < ϵ
That is, ∃x and y satisfying equations:
p
x2 + y 2 < δ

x = −y
However these x and y cannot satisfy |f (x, y) − 1| < ϵ as f (x, y) would be undefined.

Hence it is not possible to assign a limit.

Alternatively

Consider the path defined by

P1 (t) = (t, 0)

lim P (t) = (0, 0)

t→0

sin(t · 0)
f (P1 (t)) =
t+0
 sin(0)
f (P1 (t)) =
t

f (P1 (t)) = 0

lim f (P1 (t)) = 0

t→0
Now consider the path defined by:

P2 (t) = t, t2 − t

lim P2 (t) = (0, 0)

t→0

sin(t (t2 − t))

f (P2 (t)) =
t + t2 − t
sin(t (t2 − t))
f (P2 (t)) =
t2
sin(t (t2 − t))
lim f (P2 (t)) = lim
t→0 t→0 t2
As limits of both the numerator and denominator exist and both are continuous on the reals, conditions
for lhopitals rule are met. By its application:

sin(t (t2 − t)) cos(t (t2 − t)) · (3t2 − 2t)

lim = lim
t→0 t2 t→0 2t
By distributivity of multiplication and linearity of limits

cos(t (t2 − t))3t cos(t (t2 − t))

= lim − lim
t→0 2 t→0 1
By direct substitution,

cos(t (t2 − t))3t cos(t (t2 − t))

lim − lim =0−1
t→0 2 t→0 1
Hence,

cos(t (t2 − t)) · (3t2 − 2t)

lim = −1
t→0 2t
And,

sin(t (t2 − t))

lim = −1
t→0 t2

lim f (P2 (t)) = −1

t→0
Therefore as

lim f (P1 (t)) ̸= lim f (P2 (t))

t→0 t→0
lim(x,y)→(0,0) f (x, y) does not exist.
1.3 c
Based on the produced graph, if it were possible to assign a limit to g(x, y), it would appear that
lim(x,y)→(0,0) g(x, y) = 1. To verify, it must be proved that:
p
∀ϵ > 0, ∃δ > 0 such that if 0 < |(x, y) − (0, 0)| < δ ⇔ 0 < x2 + y 2 < δ

|g(x, y) − 1| < ϵ

sin(xy)
−1 <ϵ
xy
p
From 0 < x2 + y 2 < δ, it follows that:

√ p
|x| ⇔ x2 ≤ x2 + y 2

|x| ≤ δ ⇒ |x|2 ≤ δ 2

p p
|y| ⇔ y 2 ≤ x2 + y 2

|y| ≤ δ ⇒ |y|2 ≤ δ 2

p
|xy| ⇔ (xy)2 ≤ max(|x|, |y|)2 ≤ δ 2

|xy|2 ≤ δ 4

p
|xy|2 ⇔ ( (xy)2 )2 ⇒ (xy)2 ≤ δ 4

(xy)2 ≤ δ 4
Given that ∀t ∈ R:

sin(t)
− 1 < t2
t

By the closure of multiplication over R, for x ∈ R and y ∈ R, xy ∈ R. Therefore:

sin(xy)
− 1 < (xy)2
xy

Moreover,

(xy)2 ≤ δ 4
Hence,

sin(xy)
− 1 < δ4
xy
1
Set δ = ϵ 4 . This ensures that:

sin(xy)
−1 <ϵ
xy

1 p
Hence it is proved that ∀ϵ > 0, ∃δ := ϵ 4 such that if 0 < x2 + y 2 < δ, it follows that:

|g(x, y) − 1| < ϵ
Therefore lim(x,y)→(0,0) g(x, y) = 1

2 Question 2.
2.1 a
L : Rn → R is continuous at a = (a1 , a2 , a3 , ...an )
ηi : R → R

lim ηi (t) = ai ∀i ∈ [1, n]

t→k

To show:

lim L(η1 (t), η2 (t), η3 (t), ...ηn (t)) = L(a)

t→k

As limx→a L(x) = L(a),

v
u n
uX
∀ϵ > 0, ∃δ > 0 such that if 0 < |x − a| < δ ⇔ 0 < t (xi − ai )2 < δ
i=1

0 < |L(x) − L(a)| < ϵ

As limt→k ηi (t) = ai ∀i ∈ [1, n],

∀ϵi > 0, ∃δi > 0 such that if 0 < |t − k| < δi

0 < |ηi (t) − ai | < ϵi

Set
δ
ϵi =
n
Pick δη such that δη = min (δi ∀i ∈ [1, n])
Therefore, ∀ηi (t), if 0 < |t − k| < δη :
δ p δ
0 < |ηi (t) − ai | < ⇔ 0 < (ηi (t) − ai )2 <
n n

Sum over all i. This holds as |ηi (t) − ai | < nδ ∀i from previous condition. Morover, as |ηi − ai | > 0, all
of the below expressions are > 0.
n
X δ
0< |ηi (t) − ai | < n ·
i=1
n
n
X
0< |ηi (t) − ai | < δ
i=1

Iterative use of triangle inequality

n
X n
X
| (ηi (t) − ai )| < |ηi (t) − ai |
i=1 i=1
v !2
n
u n
X u X
0<| (ηi (t) − ai )| < δ ⇔ 0 < t (ηi (t) − ai ) < δ
i=1 i=1

As sum of squares ≤ square of sum:

n n
!2
X X
(ηi (t) − ai )2 ≤ (ηi (t) − ai )
i=1 i=1
v v !2
u n u n
u X
uX
t (ηi (t) − ai )2 ≤ t (ηi (t) − ai )
i=1 i=1

Therefore
v
u n
uX
0 < t (ηi (t) − ai )2 < δ
i=1

It has hence been shown that ∃δη such that if 0 < |t − k| < δη :
δ
|ηi (t) − ai | < ∀i ∈ [1, n]
n
v
u n
uX
0 < t (ηi (t) − ai )2 < δ
i=1

Assemble η(t) ∈ Rn such that η = (η1 (t), η2 (t), η3 (t), ...ηn (t)).

From previous inequality, if 0 < |t − k| < δη :

 v
u n
uX
0 < t (ηi (t) − ai )2 < δ ⇔ 0 < |η(t) − a| < δ
i=1

The definition of the limit of limx→a L(x) = L(a) required x = (x1 , x2 , x3 , ...xn ) ∈ Rn such that:
v
u n
uX
0 < |x − a| < δ ⇔ 0 < t (xi − ai )2 < δ
i=1

Hence, x can be set to η(t). Therefore, the limit definition of limx→a L(x) = L(a) can be re-written
as:

∀ϵ > 0, ∃δ > 0, ∃δη > 0 such that if 0 < |t − k| < δη it follows that
v
u n
uX
0 < |η(t) − a| < δ ⇔ 0 < t (ηi (t) − ai )2 < δ and,
i=1

0 < |L(η(t)) − L(a)| < ϵ

Which should be equivalent to saying that

lim L(η1 (t), η2 (t), η3 (t), ...ηn (t)) = L(a)

t→k

2.2 b
The statement proved is that: For f : Rn → R, if limit limx→a f (x) to exist, it is necessary that every
path in Rn approaching a gives the same limiting value.

3 Question 3.
f (a, b) = x4 + 2y 3

P = (1, 1, f (1, 1))

P = (1, 1, 3)

3.1 a
T (x, y) = Dx f (1, 1)(x − 1) + Dy f (1, 1)(y − 1) + f (1, 1)

Dx f (x, y) = 4x3 ⇒ Dx f (1, 1) = 4

Dy f (x, y) = 6y 2 ⇒ Dy f (1, 1) = 6

T (x, y) = 4(x − 1) + 6(y − 1) + 3

3.2 b
T (1.5, 1.5) = 4(1.5 − 1) + 6(1.5 − 1) + 3

T (1.5, 1.5) = 4 · 0.5 + 6 · 0.5 + 3

T (1.5, 1.5) = 2 + 3 + 3

T (1.5, 1.5) = 8

3.3 c
syms x y
f(x,y) = x^4 + 2*y^3;
P = f(1,1);
T(x,y) = 4*(x-1)+6*(y-1)+3;
fsurf(f,’FaceColor’,’r’)
hold on
fsurf(P,’FaceColor’,’g’)
hold on
fsurf(T,’FaceColor’,’b’)
xlim([1 1.5])
ylim([1 1.5])
zlim([2 12])
Or alternatively using matlab
syms x y
f(x,y) = x^4 + 2*y^3;
P = f(1,1);
T(x,y) = 4*(x-1)+6*(y-1)+3;
fsurf(f,’FaceColor’,’r’)
hold on
fsurf(P,’FaceColor’,’g’)
hold on
fsurf(T,’FaceColor’,’b’)
hold on
plot3([1,1.5],[1,1],[3,3],’Color’, [0.7 0.7 0.7], LineWidth=5)
plot3([1,1],[1,1.5],[3,3],’-y’, LineWidth=5)
plot3([1.5,1.5],[1.5,1.5],[3,3+4*(1.5-1)],’-k’, LineWidth=5)
plot3([1.5,1.5],[1.5,1.5],[3+4*(1.5-1),3+4*(1.5-1)+6*(1.5-1)],’Color’,’#EDB120’, LineWidth=5
plot3([1.4,1.4],[1.5,1.5],[3,T(1.5,1.5)],’-c’, LineWidth=5)
plot3([1.5,1.5],[1.4,1.4],[3,3.5],’Color’, [0.7 0.7 0.7], LineWidth=5)
plot3([1.5,1.5],[1.4,1.4],[3+4*(1.5-1),3+4*(1.5-1)+0.5],’-y’, LineWidth=5)
xlim([1 1.5])
ylim([1 1.5])
zlim([2 12])
Where
Gray/[0.7,0.7,0.7] = A
Yellow = B
Black = C
#EDB120 = D
Cyan/’-c’ = E