2017-2018 MATH1030

Math1030 D/E/F Midterm 2 Solution

For reference only
Name (English) (Chinese)

Student ID: Signature:

Section: D/E/F

• There are 4 pages (including this page). Please check the number of pages.
• There are 8 questions. You have to answer all the questions. The total score is 90.
• Time allowed: 120 mins.
• Please show your steps unless otherwise stated.
• No calculators allowed.
• You can use any results from the lectures (Lecture 1 to Lecture 13), notes and
sections of the textbook we covered. For other results, please give detailed ex-
planations. In particular, you cannot use the theory of determinants or
inverses of matrix
• Good luck.

1
1. (12 pts) Let       

 0 1 1 

 1 1 1
S = u1 =   , u2 =   , u3 =   .
     

 2 2 3 

 −1 −1 0 
(a) Find real numbers α1 , α2 , α3 , α4 such that
 
v1
v2 
v= 
v3 
v4

is in hSi if and only if

α1 v1 + α2 v2 + α3 v3 + α4 v4 = 0.

6 R4 by finding a vector v in R4 but not in hSi.

(b) Show that hSi =
Answer.
(a) Consider the matrix A = [u1 |u2 |u3 ], then v is in hSi if and only if LS(A, v) is
consistent. The augmented matrix is:

   
0 1 1 v1 1 0 0 v2 − v1
 1 1 1 v2  
 (4pts) −→  0 1 0 v1 + 2v2 − v3 
[A|v] =   (4pts)
 2 2 3 v3   0 0 1 −2v2 + v3 
−1 −1 0 v4 0 0 0 3v2 − v3 + v4

LS(A, v) is consistent if and only if the last column is not a pivot column.
Therefore 3v2 − v3 + v4 = 0. (2 pts)
 
0
0 4
0, then 3v2 − v3 + v4 6= 0. Therefore v in R but not in hSi and
(b) Let v =  
1
6 R4 . (2 pts)
hence hSi =

2. (13 pts) Let  
1 2 0 2 −1 1
 2 4 1 7 −2 1 
A= 
 −3 −6 0 −6 4 −3 
4 8 −2 2 1 6
For this question, no steps are required.

2
(a) Let B be the RREF of A, find B. Hint: [B]14 = 2, [B]26 = −1.
(b) From the previous part, find a linearly independent set S, such that hSi =
N (A).
(c) Determine which of the following vectors
0 0
   
2 1
1 1
   
u =  ,v =  .
−1 −1
0 0
−2 1

are in the N (A). If it is in the N (A), express it as a linear combination of S.

(d) Let b ∈ R4 such that
1
 
1
0
 
x= 
0
1
0
is a solution of LS(A, b). Write down the solution set of LS(A, b) using part
(b).
Answer.
(a)  
1 2 0 2 0 1
 0
 0 1 3 0 −1 
.
 0 0 0 0 1 0 
0 0 0 0 0 0
(Remark: correct answer (3 pts); but each mismatch entry (-1 pts).)
(b)
−2 −2 −1
     
1 0 0
0 −3 1
     
z1 =   , z2 =   , z3 =  
0 1 0
0 0 0
0 0 1
S = {z1 , z2 , z3 }
(Remark: correct answer (3 pts); but each minor mistake (-1 pts).)
(c) ‘Yes’ for u and ‘No’ for v because u = 2z1 − z2 − 2z3 ∈ hSi = N (A) but
Av 6= 0.

3
 (Remark: correct yes/no (1+1 pts) and correct combination (3 pts); but each
minor mistake (-1 pts).)
(d) The solution set is given by x + N (A).
1 −2 −2 −1
        
 
1 1 0 0

 


         

0 0 −3 1
        
x + N (A) =   + α   + β   + γ   α, β, γ ∈ R

 0 0 1 0 


 1 0 0 0 


 

0 0 0 1

(Remark: correct answer (2 pts); but each minor mistakes (-1 pts).)

3. (10 pts) Let a be a fixed real number.
       
1 2 a 3
1 0 0 −1
v1 = 0 , v2 = 1 , v3 = 0 , v4 =  2 
      
0 0 1 3

(a) Show that {v1 , v2 , v3 } is linearly independent.

(b) Find all a’s such that {v1 , v2 , v3 , v4 } is linearly dependent.
(c) Suppose a is different from the answer obtained in the previous part, what
can you say about the square matrix [v1 |v2 |v3 |v4 ]? What is h{v1 , v2 , v3 , v4 }i?
Give reasons.

Answer.
(a) Let α1 v1 + α2 v2 + α3 v3 = 0. That is,
       
α1 2α2 aα3 0
α1   0   0  0
 +
 0   α2  +  0  = 0 .
    
0 0 α3 0

By considering the 2nd, 3rd and 4th entries, α1 = α2 = α3 = 0. (4 pts)

(b) Suppose {v1 , v2 , v3 , v4 } is linearly dependent. From definition, there exist αi ∈
R, 1 ≤ i ≤ 4, not all zero, such that α1 v1 + α2 v2 + α3 v3 + α4 v4 = 0. That is,
         
α1 2α2 aα3 3α4 0
α1   0   0  −α4  0
 +
 0   α2  +  0  +  2α4  = 0 .
      
0 0 α3 3α4 0

4
 From the 2nd, 3rd and 4th entries, we obtain:
α1 = α4
α2 = −2α4
α3 = −3α4 .
Consider the first entry together with the above substitution, we have
α4 − 4α4 − 3aα4 + 3α4 = 0.
Which is just
aα4 = 0.
If α4 = 0, the above implies that αi = 0 for all i, which contradicts the
assumption. Therefore we have a = 0. (4 pts)
Remark : Marks will be deducted if one assumes α4 = 1 without detail expla-
nation.
(c) If a is different from 0, {v1 , v2 , v3 , v4 } is linearly independent. Therefore
[v1 |v2 |v3 |v4 ] is nonsingular. Since each vi is in R4 , and {v1 , v2 , v3 , v4 } contain-
ing exactly 4 vectors is linearly independent, h{v1 , v2 , v3 , v4 }i = R4 .(2 pts)
Remark : 1 mark for nonsingular or any equivalent correct answers, 1 mark for
the span.

4. (14 pts) Determine which of the following subsets are subspaces of R4 :
  
x
  1

 

x2 

(a) W1 =  ∈ R4 x1 + x2 = x3 + x4 and x1 − x2 = x3 − x4 ;


x3  

 x 
4
  
x
  1

 

x2 

4
(b) W2 = 
 x 3 ∈R x1 + x2 = x3 + x4 or x1 − x2 = x3 − x4 .


 x 

4

Justify your answers (give a proof when it is a subspace, and a concrete numerical
counterexample when it is not).
Answer.
(a) Note that 0 + 0 = 0 + 0 and 0 − 0 = 0 − 0. Thus,
 
0
0
  ∈ W1
0
0

5
 and hence W1 is non-empty. (1 pts)

Method 1: Suppose α ∈ R and

   
u1 v1
u2  v2 
u3  , v = v3  ∈ W1 .
u=   
u4 v4
Then
(αu1 + v1 ) + (αu2 + v2 ) = α(u1 + u2 ) + (v1 + v2 ) = α(u3 + u4 ) + (v3 + v4 )
= (αu3 + v3 ) + (αu4 + v4 ); (3pts)
(αu1 + v1 ) − (αu2 + v2 ) = α(u1 − u2 ) + (v1 − v2 ) = α(u3 − u4 ) + (v3 − v4 )
= (αu3 + v3 ) − (αu4 + v4 ).(3pts)
Thus,
 
αu1 + v1
αu2 + v2 
αu3 + v3  ∈ W1 .
αu + v =  
αu4 + v4
Method 2: Suppose
   
u1 v1
u2  v2 
u3  , v = v3  ∈ W1 .
u=   
u4 v4
Then
(u1 + v1 ) + (u2 + v2 ) = (u1 + u2 ) + (v1 + v2 ) = (u3 + u4 ) + (v3 + v4 )
= (u3 + v3 ) + (u4 + v4 );
(u1 + v1 ) − (u2 + v2 ) = (u1 − u2 ) + (v1 − v2 ) = (u3 − u4 ) + (v3 − v4 )
= (u3 + v3 ) − (u4 + v4 ).(3pts)
Thus,
 
u1 + v1
u2 + v2 
u3 + v3  ∈ W1 .
u+v = 
u4 + v4
Suppose α ∈ R and
 
u1
u2 
u3  ∈ W1 .
u= 
u4

6
 Then

αu1 + αu2 = α(u1 + u2 ) = α(u3 + u4 ) = αu3 + αu4 ;

αu1 − αu2 = α(u1 − u2 ) = α(u3 − u4 ) = αu3 − αu4 .(3pts)

Thus,
 
αu1
αu2 
αu3  ∈ W1 .
αu =  
αu4

In conclusion, W1 is a subspace of R4 . (1 pts)

(One mark for showing W1 is non-empty.
Six marks for each of the two methods.
One mark for the final conclusion given that the previous explanation is correct.
Only writing ‘similarly’ to skip calculations is not acceptable)
(b) Let
  
1 1
−1 1
u=
0
 and v = 
0 .

0 0

Since 1 + (−1) = 0 = 0 + 0 and 1 − 1 = 0 = 0 − 0, u, v ∈ W2 . (2 pts)

However, 2 + 0 = 2 6= 0 = 0 + 0 and 2 − 0 = 2 6= 0 = 0 − 0. So
 
2
0
u+v = 0 6∈ W2 .(3pts)

0

In conclusion, W2 is not a subspace of R4 . (1 pts)

(Two marks for giving a correct counterexample of u, v.
Three marks for showing that u + v 6∈ W2 .
One mark for the final conclusion given that the previous explanation is correct.
Giving non-concrete / non-numerable counterexamples gains no mark.)

5. (10 pts) For this question, you can use the following fact (you don’t need to prove
the fact):
Fact Let A and B be two row equivalent matrices. Let Ai (resp. Bi ) be the i-th
column of A (resp. B).
α1 Ai1 + α2 Ai2 + · · · + αk Aik = Aj

7
 if and only if
α1 Bi1 + α2 Bi2 + · · · + αk Bik = Bj .
Suppose that the matrices
   
2 a e 2 4 1 2 0 h 0
 b 2 3 f 1   0 0 1 −1 0 
A=  2 c
 and B =  
g 0 0   0 0 0 0 1 
d 6 7 5 9 0 0 0 0 0

are row equivalent.

(a) Express A2 as a linear combination of {A1 } and hence find a, b, c, d.

(b) Express A4 as a linear combination of {A1 , A3 } in terms of h, and hence find
e, f, g, h.
Answer.
(a) From the Fact, As B2 = 2B1 , A2 = 2A1 .(1 pts) Therefore,
 

 a = 2 · 2 
 a=4
2=2·b b=1
 
−→ (2pts)

 c=2·2 
 c=4
6 = 2 · d d = 3

(b) From the Fact, As B4 = hB1 − B3 , A4 = hA1 − A3 .(1 pts) Therefore,

 

 2 = h · 2 − e 
 e=6
f =h·1−3 f =1
 
(4pts) −→ (4pts)

 0=h·2−g 
 g=8
5 = h · 3 − 7 h = 4


6. (13 pts) Let S = {v1 , v2 , v3 , v4 } be a subset of Rm . Suppose
v1 + 3v3 − v4 = 0.

(a) Let α1 , α2 , α3 , α4 be real numbers. Express

α1 v1 + α2 v2 + α3 v3 + α4 v4
as a linear combination of v2 , v3 , v4 and hence show that hSi = h{v2 , v3 , v4 }i.
(b) Let m = 4. Is it always true that hSi = h{v1 , v3 , v4 }i? If it is true, prove it, if
it is not true, give a counter example.
Answer.

8
 (a)
v1 + 3v3 − v4 = 0
is equivalent to
v1 = −3v3 + v4 .
Consider
α1 v1 + α2 v2 + α3 v3 + α4 v4
= α1 (−3v3 + v4 ) + α2 v2 + α3 v3 + α4 v4
= α2 v2 + (α3 − 3α1 )v3 + (α1 + α4 )v4 .

Let v ∈ hSi.
v = α1 v1 + α2 v2 + α3 v3 + α4 v4 ,
for some αi ∈ R, 1 ≤ i ≤ 4. From the above, v can be expressed as
linear combination of v2 , v3 , v4 . Therefore v ∈ h{v2 , v3 , v4 }i, and we have
hSi ⊂ h{v2 , v3 , v4 }i.

Let v ∈ h{v2 , v3 , v4 }i,

v = α2 v2 + α3 v3 + α4 v4
= 0v1 + α2 v2 + α3 v3 + α4 v4
for some αi ∈ R, 2 ≤ i ≤ 4. v can be expressed as linear combination of
v1 , v2 , v3 , v4 , therefore h{v2 , v3 , v4 }i ⊂ hSi. In conclusion, hSi = h{v2 , v3 , v4 }i.
(7 pts)
Remark : 3 marks for for the expression, 4 marks for showing hSi = h{v2 , v3 , v4 }i.
Careful explanations are needed for the later part.
     
1 4 0
0 0 1
(b) No, consider v1 = v3 =  0 , v4 = 0, v2 = 0. We can see that v1 + 3v3 −
    
0 0 0
   
0 α
1 0
v4 = 0, and v2 ∈ hSi. However v2 =  0 ∈
 / {  | α ∈ R} = h{v1 , v3 , v4 }i.
0
0 0
(6 pts)
Remark : 3 marks for valid counter example, 3 marks for explanations.

7. (6 pts) Suppose S = {u1 , u2 , u3 , u4 } is in Rm . Suppose v can be expressed
as linear combination of S in two distinct ways, i.e., there exist real numbers
α1 , α2 , α3 , α4 , β1 , β2 , β3 , β4 such that
v = α1 u1 + α2 u2 + α3 u3 + α4 u4 = β1 u1 + β2 u2 + β3 u3 + β4 u4

9
 and (α1 , α2 , α3 , α4 ) 6= (β1 , β2 , β3 , β4 ). Show that S is linearly dependent.
Answer.

v = α1 u1 + α2 u2 + α3 u3 + α4 u4
v = β1 u1 + β2 u2 + β3 u3 + β4 u4
• Given the above two distinct linear combinations, we can subtract them to get
the following.
(α1 − β1 )u1 + (α2 − β2 )u2 + (α3 − β3 )u3 + (α4 − β4 )u4 = 0

• Since (α1 , α2 , α3 , α4 ) and (β1 , β2 , β3 , β4 ) are distinct, (αi − βi )’s are not all zero.
More precisely, there exists some index 1 ≤ k ≤ 4 such that (αk − βk ) 6= 0.
Hence, {u1 , u2 , u3 , u4 } is linearly dependent.
(Remark: correct rearrangment and argument (3+3 pts); but with minor
mistakes (-1 pts), such as missing ideas like ‘not all zero’, ‘exist some’, etc; or,
picking v = 0 (-6 pts).)

8. (12 pts) Let n be an integer ≥ 2. Let S = {u1 , u2 , u3 , . . . , un } be a subset of
distinct nonzero vectors in Rm with a nontrivial relation of linear dependence
α1 u1 + α2 u2 + · · · + αn un = 0,
where αi , i = 1, . . . , n are real numbers.
Consider the following three properties:
Property A There exist n − 1 distinct vectors of S that are linearly independent.
Property B Any n − 1 distinct vectors of S are linearly independent.
Property C Suppose that
β1 u1 + β2 u2 + · · · + βn un = 0,
where βi , i = 1, . . . , n, are real numbers. Then there exists a real number t such
that
βi = tαi
for i = 1, . . . , n.
(a) Show that if Property C is true, then Property A is true.
(b) Let n = 3, m = 4. Find an explicit numerical example to show that even if
Property C is true, Property B may not be true.
Answer.
(a) Prove by contradiction, suppose Property A is incorrect. That is any n − 1
vectors in S is linearly dependent.

10
Next because
α1 u1 + α2 u2 + · · · + αn un = 0,
is a nontrivial relation. There exists i such that αi 6= 0. (2pts) Then we have
nontrivial relation:
β1 u1 + · · · βi−1 ui−1 + βi+1 ui+1 βn un = 0 (1)

Set βi = 0, we have
β1 u1 + · · · βi−1 ui−1 + βi ui βi+1 ui+1 βn un = 0

By Property C, t = α0i = 0. Thus βk = αk t = 0 for all i = 1, · · · , k. This contradicts

to the assumption that the relation (1) is nontrivial. (4pts)

     
1 0 0
 0   1   2 
(b) Set u1 =   0 , u2 =  0 , u3 =  0  and A = [u1 |u2 |u3 ] (2pts). The
    
0 0 0
* +
0
solotion set of Ax = 0 is  2  , thus Property C is satisfied. (2pts) However,
1
{u2 , u3 } are linearly dependent, Property B is not true. (2pts) 

—————— The end ——————

11