Linear Algebra (Matrices)

Problem Set 10 12 December, 2022

This weeks problem set contains no new material. Instead it consists of a few past exam questions to test
your understanding and help you prepare for the exam in the new year.
Exam 2012 Question 6 - Eigenvalues

6. Consider the 2 × 2 matrix,  

2 −3
B= .
2 −5

(i) Calculate det(B ), the determinant of the matrix B .

det(B ) = −10 + 6 = −4

(ii) Find the two eigenvalues λ1 > λ2 of the matrix B .

2 − λ −3
charactenstic equation p(λ ) = =0
2 −5 − λ
(2 − λ )(−5 − λ ) + 6
<0
−10 + 3λ + λ 2 + 6 =0
λ 2 + 3λ − 4 =0
(λ + 4)(λ − 1) =0
Eigenvalues are λ1 = 1, λ2 = −4.
 
3
(iii) Show that x1 = is an eigenvector of the matrix B with eigenvalue λ1 .
1
    
2 −3 3 3
B x1 = = = λ1 x1
2 −5 1 1
Indeed x1 is an eigenvector of B with eigenvalue λ1 = 1.

(iv) Find the eigenvector x2 associated with the smallest eigenvalue λ2 .

For λ2 = −4,
      
6 −3 x 0 1
= ⇒ 2x − y = 0 ⇒ y = 2x ⇒ x2 =
2 −1 y 0 2

(v) Let S = (x1 x2 ) denote the 2 × 2 matrix of eigenvectors. Calculate S −1 B S explicitly.

   
31 −1 1 2 −1
S= S =5
12 −1 3
      
1 31 2 −1 1 50 10
check 5 =5 =
12 −1 3 05 01
   
1 2 −1 2 −3 31
So S −1 B S =
5 −1 3 2 −5 12
  
1 2 −1 3 −4
=
5 −1 3 1 −8
     
1 5 0 10 λ1 0
= = =
5 0 −20 0 −4 0 λ2
 (vi) Why does the inverse B −1 exist?

B −1 exists because det(B ) ̸= 0

(vii) What are the two eigenvalues of B −1 ?

since B xi = λi xi
−1
and B B xi = B −1 (λi xi ) = λi B −1 (xi )
1
⇒ B−1 x1 = xi
λ
Hence, the eigenvalues of B−1 are λ = 1, − 41

(viii) What are the two eigenvectors of B −1 ?

   
3 1
and by above xi are also eigenvectors of B −1 . So eigenvectors are and
1 2

Exam 2018 Question 3 - Linear Dependence:

1. (i) Explain the difference between linear dependence and independence of a set of n vectors v1 , v2 , ..., vn
in Rn . How is the determinant of this set of vectors related to their being linearly dependent or
independent?
For a linearly dependent set of vectors any one vector can be written as a superposition of all
others, e.g. v1 = c11 (c2 v2 + c3 v3 + . . .). Linearly dependent vectors cannot.
A linearly dependent set will have the determinant of the matrix formed of the column vectors
equal to zero.

(ii) A set of simultaneous linear equations A x = b can be written as

Ai xi + A2 x2 . . . An xn = b,

in which A1 x1 is the i th column of A (i = 1, 2, ..., n) and xi is the i th entry of the column vector x
(i = 1, 2, ..., n).
(a) Referring to your answer in (i), under what conditions on the Ai is there a unique solution to
this equation?
There is a unique solution if det(A ) ̸= 0 and b ̸= 0 or in other words if the Ai are linearly
independent.
(b) Suppose b = 0. Again referring to your answer in (i), under what conditions on the Ai are
there non-trivial solutions to A x = 0?
In this case, det(A ) = 0 or the only solution is the trivial one with all xi = 0.
(c) Suppose that the Ai are linearly dependent. What determines the nature of the solutions to
A x = b?
The nature of the solutions depends on the level of dependence. The solutions will span a Rn
subspace where n is the number of independent vectors. If there is two, all solutions lie on a
plane, if only one on a line etc.
 (iii) Consider the following system of 3 equations in 3 unknowns:
x − z = b1 ,
y + z = b2 ,
x + y = b3 .

(a) Use Gaussian elimination to determine the conditions on b1 , b2 , and b3 that yield (i) no
solution and (ii) an infinity of solutions of this system.
By Gaussian elimination:
     
1 0 −1 b1 1 0 −1 b1 1 0 −1 b1
 0 1 1 b2  →  0 1 1 b2  →  0 1 1 b2 
1 1 0 b3 0 1 1 b3 − b1 0 0 0 b3 − b1 − b2
These equations are only consistent if b3 − b1 − b2 = 0 in which case there is an infinity of
solutions.
Otherwise there are no solutions.
(b) Determine an expression for the solutions when there is an infinity of solutions to this system.
For the case the equations are consistent, then we can  = λ . 
set z  
  x b1 1
x − z = b1 x = b1 + λ  y  =  b2  + λ  −1 
Reading the rows:
y + z = b2 y = b2 − λ
z 0 1

Exam 2012 Question 5 - Planes:

1. Consider the three vectors a = î − 2 ĵ − 3 k̂, b = î − 4 ĵ − 13 k̂ and c = −3 î + 5 ĵ + 4 k̂ where î, ĵ and k̂

denote the natural basis vectors for R3 .

(i) Calculate the lengths |a|, |b| of the vectors a and b.

√ √
a = 1 + 4 + 9 = 14
√ √
b = 1 + 16 + 169 = 186

(ii) Calculate the dot (scalar) product a · b.

   
1 1
a · b =  −2  ·  −4  = (1 + 8 + 39) = 48
−3 −13

(iii) From the sign of the dot product a · b, what can you conclude about the angle θ (0 ≤ θ ≤ π)
between the two vectors a and b?

since a · b = |a||b| cos θ > 0 0 < θ < π/2

(iv) Find the angle θ (0 ≤ θ ≤ π) the two vectors a and b.

 
48 −1 48
Then cos θ = p √ ⇒ θ = cos √ √ = 0.346 rad = 19.8◦
14 186 14 186

(v) Calculate the area of the parallelogram spanned by the two vectors a and c.
 î ĵ k̂ î ĵ k̂ î ĵ k̂
|a × c| = 1 −2 −3 = 1 −2 −3 = 1 0 7 = 7 î + 5 ĵ − k̂
−3 5 4 0 −1 −5 0 −1 −5
√ √ √
and |A| = |a × c| = 49 + 25 + 1 = 75 = 5 3

Consider three planes in R3 specified by the equations,

x − 2y − 3z = 2, (1a)
x − 4y − 13z = 14, (1b)
−3x + 5y + 4z = 0. (1c)

(vi) Let A denote the matrix of coefficients for this system of three linear equations in three unknowns.
Calculate det(A ), the determinant of A .

1 −2 −3 1 0 0
22
det(A) = 1 −4 −13 = 1 −2 −10 =5 =0
11
−3 5 4 −3 −1 −5

(vii) Does the system of linear equations (1a), (1b) and (1c) have a unique solution?
No, the equations do not have a unique solution.

(viii) Using Gaussian elimination or otherwise, show that the solution r = (x, y, z) to equations (1a), (1b)
and (1c) can be written in the form

r = r0 + µd, µ ∈R

and identify possible values for r0 and d.

   
1 −2 −3 2 1 −2 −3 2 1 −2 −3 2 1 0 7 −10
1 −4 −13 14 → 0 −2 −10 12 →  0 1 5 −6  →  0 1 5 −6 
−3 5 4 0 0 −1 −5 6 0 1 5 −6 000 0
R2 → R2 − R1 R2 → R2 ÷ −2 R1 → R1 + 2R2
R3 → R3 + 3R1 R3 → R3 × −1 R3 → R3 − R2
Here z is arbitrary, so let z = µ
so soln is x = −10 − 7z = −10 − 7µ
y = −6 − 5z = −6 − 5µ
z =µ
     
x −10 −7
⇒  y  =  −6  + µ  −5 
z 0 1

Exam 2020 Question 1

(You can find the question and solutions for this paper elsewhere but are included here for convenience.)

1. (a) A particle starts at position r0 = (0, 1, 0) with a velocity v = (2, 1, 3). It collides with a wall whose
position is given by the equation x + y + z = 4.
You may assume throughout this question that all physical quantities are given in dimensionless
units, and there are no external forces acting on the particle.
 i) Give an expression for the trajectory of the particle before the collision.
We have, r = r0 + vt
   
0 2
r =  1  +  1 t
0 3

ii) Find the position at which the particle strikes the wall.

From above, x = 2t, y = 1 + t, z = 3t

in x + y + z = 4, ⇒ 2t + (1 + t) + 3t = 4
⇒ 6t = 3
⇒ t= 1/2    
0 2 1
So rcoll =  1  + 21  1  =  3/2 
0 3 3/2

iii) Find the velocity of the particle after the collision with the wall, assuming the collision is
elastic.
v⊥
After collision v|| is unchanged
v||
n̂ but v⊥ is ‘negated’
v||
v⊥ = vproj = ( n̂ · v) n̂
v⊥
   
1 1
1  
n= 1
 → n̂ = 3 1
√
1 1
         
1 2  1 1 2
1 
 1  ·  3  √1  1  = 1 {6} ·  1  =  2 
vproj = √
3 1  3 3
1 1 1 2
     
2 2 0
Then v|| =  1  −  2  =  −1 
3 2 1
 
−2
After collision, v⊥ ′ =  −2 
−2
     
−2 0 −2
and so vafter = v⊥ ′ + v|| =  −2  +  −1  =  −3 
−2 1 −1
(check - energy unchanged)
(NB this could also be found by finding the transformation matrix for the reflection from
x + y + z = 0 but finding the reflection matrix in this case is not trivial.)
(b) For the matrix,  
1 1+a
A= ,
1+a 1
i) find the eigenvalues and eigenvectors of A .
 1−λ 1+a
First to find eigenvalues λi . we want p(x) = =0
1+a 1−λ
λ 2 − 2λ + 1 − 12 + 2a + a2 = 0


→
→ λ 2 − 2λ − 2a√− a2 = 0
→ λ = 21 (+2
√ ± 4 + 8a + 4a )
2
p
= 1 ± a2 + 2a + 1 = 1 ± (1 + a)2 = 1 ± (1 + a)
→ λ = −a, 2 + a
for λ = −a,  
1+a 1+a x
⇒ =0
1+a 1+a y  
1
⇒ x+y = 0 → y = −x → v1 =
−1
for λ = 2 + a,  
−(1 + a) 1 + a x
⇒ =0
1 + a −(1 + a) y  
1
⇒ x−y = 0 → y=x → v2 =
1
ii) Hence, find A 10 .

We have A 10 = S Λ10 S −1
where S is the matrix of eigenvectors and Λ is the diagonal matrix of the eigenvalues.
So matrixof eigenvectors,
  
1 1 −1 1 1 −1
S= and S = 2
−1 1 1 1
check,
1
− 12
       
1 1 1 0 1 1 1 0 1 1 1 0 1 0 2
→ → →
−1 1 0 1 0 2 1 1 0 1 12 12 0 1 1
2
1
2
 
−a 0
and Λ=
0 2+a
and A 10 = S Λ10 S −1
  10  
1 11 −a 0 1 −1
=
2 −1 1 0 2+a 1 1
   10  
1 11 a 0 1 −1
=
2 −1 1 0 (2 + a)10 1 1
10 10
  
1 1 1 a −a
=
2 −1 1 (2 + a)10 (2 + a)10
1 (2 + a)10 + a10 (2 + a)10 − a10
 
=
2 (2 + a)10 − a10 (2 + a)10 + a10

iii) Find A 10 for a = 0.01.

for a = 0.01,
10

a10 = (0.01)10 = 10−2 = 10−20 ≈ 0
and (2 + a)10 = (2.01)10 ≈ 1076.4
1
and 2 (2 + a)10 = 538 · 2
 
538.2 +538.2
→ A10 ≈
+538.2 538.2