Mathematics I for EOR: Homework 2:

Exercise 1
Given is the set A = {(x, y) ∈ R2 | x2 + y 2 ≤ 1}. Prove that A is a compact set.
a) Do this by directly showing that it is bounded and closed using the definition of closure.
b) Do this by proving that the complement of A is open.
HINT: Use the triangle-inequality k~ak = k~a − ~q + ~qk ≤ k~a − ~qk + k~qk to work out the algebra.

a) Bounded means that there is some number b for which we have that k~ak ≤ b for all ~a ∈ A.
So for some arbitrary ~a = (x, y), we have that
p √
k~ak = x2 + y 2 ≤ 1 = 1 = b

because of the definition of A. So A is bounded.

1 point for correctly applying the definition of bounded.

For the definition of closure: consider any convergent sequence ~an in A. Define ~a = limn→∞ ~an .
Then, by definition ~a is a point of closure of A. Hence, to prove that A is closed, we must prove
that ~a ∈ A.
1 point for writing down a conclusion like this.

Decompose ~an and ~a into components: ~an = (xn , yn ) and ~a = (x, y). Then ~a = limn→∞ ~an if
and only if x = limn→∞ xn and y = limn→∞ yn .
1 point for writing down something like this which uses and quotes the theorem that a sequence
converges if and only if its components converge.

Then, we can consider the sequence zn = xn 2 + yn 2 . Because of algebragic limit properties, we

know that z = limn→∞ zn = (limn→∞ xn )2 + (limn→∞ yn )2 = x2 + y 2 . We also know that zn ≤ 1
for all n, because ~an ∈ A for all n and the definition of A.
1 point for correctly using the algebragic limit properties of the lectures. the words ‘algebragic
limit properties’ must be there.

Then, again by algebragic limit properties, we get that z ≤ 1. Hence, x2 + y 2 ≤ 1, hence

~a = (x, y) ∈ A.
So A is bound and closed, hence compact.
q.e.d.
1 point for a proper final conclusion.

b) If A = {(x, y) ∈ R2 | x2 + y 2 ≤ 1}, then by definition, AC = {(x, y) ∈ R2 | x2 + y 2 > 1}. To

prove that AC is open, we must show that for all ~a = (x, y) ∈ AC , there exists  > 0 such that
B (~a) ⊂ AC .
1 point for correctly using the open-ball definition.
p
So consider any ~a = (x, y) ∈ AC . We can then define the positive number  = x2 + y 2 − 1 > 0.
Then, we have to prove that B (~a) ⊂ AC . We do this by considering any ~q = (u, v) ∈ B (~a)
and then proving that ~q = (u, v) ∈ AC .
1 point for correctly writing down this strategy

We will use the triangle inequality to prove this: k~ak = k~a − ~q + ~qk ≤ k~a − ~qk + k~qk
Hence, we have that k~qk ≥ k~ak − k~a − ~qk. p
Now, as ~q = (u, v) ∈ B (~a), we know by definition that k~a − ~qk <  = x2 + y 2 − 1 = k~ak − 1.

1
k~a − ~qk < k~ak − 1.
−k~a − ~qk > −k~ak + 1. So:
k~qk ≥ k~ak − k~a − ~qk > k~ak + (−k~ak + 1) = 1. So: k~qk > 1, meaning that ~q ∈ AC .
1 point for correctly working out the algebra

As we have proven that AC is open, this means that A is closed. As we already proved that A
is bound in part a), this means that A is compact.
q.e.d.
1 point for the final conclusion.

Exercise 2
Consider the following 4 vectors in R4 : ~q = (4, 3, 2, 1), ~u = (1, 1, 0, 1), ~v = (2, 0, 0, 3), and
w
~ = (1, 2, 3, 4).
~ is a valid basis for R4 .
a) Prove that {~q, ~u, ~v , w}
b) Consider the vector ~a = (4, 4, 3, 3) with respect to the standard basis {ê1 , ê2 , ê3 , ê4 }. Trans-
form ~a to the {~q, ~u, ~v , w}-basis.
~
c) Consider the vector ~b = (2, 1, 1, 2) with respect to the the {~q, ~u, ~v , w}-basis.
~ Transform ~b to
the standard basis {ê1 , ê2 , ê3 , ê4 }.

a) Any n lineary independent vectors form a basis of Rn . So to prove this, we must show that
the vectors {~q, ~u, ~v , w}
~ are lineary independent. 1 point for this conclusion.

We do this by considering the matrix

 
4 1 2 1
3 1 0 2
A = (~q ~u ~v w)
~ =
2

0 0 3
1 1 3 4

Then, we compute that

4 1 2 1
1 2 1 4 1 2
3 1 0 2
det(A) = =2· 1 0 2 −3· 3 1 0 =
2 0 0 3
1 3 4 1 1 3
1 1 3 4
   
2 1 1 2 1 2 4 2
2 · −1 · −2· − 3 · −3 · +1· =
3 4 1 3 1 3 1 3
2 · (−1 · 5 − 2 · 1) − 3 · (−3 · 1 + 1 · 10) = −35 6= 0
So A is non-singular and, hence, has maximal rank. So {~q, ~u, ~v , w}
~ are lineary independent.
1 point for a computation like this (do not count typo-miscalculations).

b) In this case, we are looking for a vector ~a = (aq , au , av , aw ), such that

aq · ~q + au · ~u + av · ~v + aw · w
~ = ~a = (4, 4, 3, 3). This means that we have to solve the linear
system      
4 1 2 1 aq 4
3 1 0 2  au  4
A · ~a = 
2 0 0 3 ·  av  = 3 = ~a
    

1 1 3 4 aw 3

2
1 point writing down the correct linear system.

Solve the system by Gaussian elimination on the augmented matrix:

 
4 1 2 1 4
3 1 0 2 4
 
2 0 0 3 3
1 1 3 4 3

Replace the second row by −3/4 times the first row plus the second row.
Replace the third row by −2/4 times the first row plus the third row.
Replace the fourth row by −1/4 times the first row plus the fourth row.
 
4 1 2 1 4
 0 0.25 −1.5 1.25 1
 
 0 −0.5 −1 2.5 1
0 0.75 2.5 3.75 2
Replace the third row by +2 times the second row plus the third row.
Replace the fourth row by −3 times the second row plus the fourth row.
 
4 1 2 1 4
 0 0.25 −1.5 1.25 1 
 
0 0 −4 5 3
0 0 7 0 −1
Replace the fourth row by +7/4 times the third row plus the fourth row.
 
4 1 2 1 4
 0 0.25 −1.5 1.25 1 
 
0 0 −4 5 3 
0 0 0 8.75 4.25
Now, apply backward substitution to compute aw = 4.25/8.75 ≈ 0.486, av ≈ −0.143, au ≈ 0.714
and aq ≈ 0.771. Hence, the transformed vector is ~a = (0.771; 0.714; −0.143; 0.486).
1 point for correctly solving the linear system (do not count typo-miscalculations).

c) This is easy. If ~b is given with respect to the {~q, ~u, ~v , w}-basis,

~ then we have with respect to
the standard basis that
~b = bq · ~q + bu · ~u + bv · ~v + bw · w
~=
         
4 1 2 1 13
3 1 0 2 11
2·2 + 1 · 0 + 1 · 0 + 2 · 3 = 10
        

1 1 3 4 14
1 point for the corect methodology and 1 point for a correct calculation.

The homework has a total of 15 points. Hence, the homework grade is 0.2 · score/15.