APL703: Assignment 2

Prapanch Nair

Due: (Wednesday) September 11, 2022.

√
1. Show that the norm ∥∥ induced by an inner product according to ∥v∥ = v · v must satisfy the
parallelogram law:
∥ u + v ∥2 + ∥ u − v ∥2 = 2∥ v ∥2 + 2∥ u ∥2
for all vectors u and v.
2. For the basis {e1 , e2 , . . . , e N } of V , the inner product of any two basis vector is:
e jk = e j · ek
. Proved that the N × N matrix [e jk ] defined by this expression is non-singular.
3. Show that each orthonormal bases is self-reciprocal. Derive a formula for obtaining the reciprocal basis
for a given “orthogonal” basis (not orthonormal).
4. We learned that a vector v can be expanded in two reciprocal bases as :
N N
v= ∑ vk ek = ∑ vk ek
k =1 k =1
| {z } | {z }
contravariant components covariant components

Now use the formula for these components, discussed in class and derive the expression for inner
product of two vectors:
N N
u·v = ∑ ∑ λ j µk e jk
j =1 k =1
(Hint: use the contraction property of the Kronecker delta, writing one vector in covariant and the
other in contravariant components)
5. Find the reciprocal basis for the following shown basis:

6. Prove that the eigenvalues of a symmetric matrix are all real. Hint: The complex conjugate can be
indicated with a bar sign. Re + iIm = Re − iIm. And the following holds:
Ax = Ax = λx = λx
T
Also for a complex matrix A, a Hermitian A H is defined as A H = A . A symmetric matrix is also a
Hermitian matrix (applicable to real matrices).

1
 7. For a 2x2 matrix, show that for an eigenvalue λ is of multiplicity k, then the corresponding eigenspace
is of dimension k.
8. Let A be a symmetric n × n matrix. Applying pre and post inner product of any eigenvector e of A
into both sides of Ae = λe and solving for λ, gives

e · Ae e T Ae
λ= = T
e·e e e
Generally, if x is any vector, not necessarily an eigenvector of A, then we can define a number:

x · Ax x T Ax
R( x ) = = T
x·x x x
which is called the Rayleigh’s quotient. If x coincides with an eigen vector, then R( x ) give the
corresponding eigenvalue. If |λ1 | ≥ |λ2 | ≥ . . . ≥ |λn |, then prove that

x · Ax
| R( x )| = ≤ | λ1 |
x·x

for a symmetric A. HINT: Since A is symmetric, it has n orthogonal eigenvectors e1 , e2 , . . . , en ,

corresponding to eigenvalues λ1 , λ2 , . . . , λn . Expand x in terms of the orthogonal basis {e1 , e2 , . . . , en }
:
x = Σnj=1 a j e j .

Verify the above conclusion for the below matrix:

 
2 0 0
A = 0 −1 −2
0 −2 −1

using the vectors x = [1, 0, 0] T , [0, 1, 0] T , [0, 0, 1] T , [1, 2, 3] T and [3, −1, 4] T , verifying that | R( x ) ≤ |λ1 | in
each case.
9. From eigenvectors of the following matrices (n × n) obtain an orthogonal basis for Rn .
(a)  
0 1 0
1 0 0
0 0 0
(b)  
0 0 0 4
0 0 4 0
 
0 4 0 0
4 0 0 0
(c)  
7 4 −4
 4 1 8 
−4 8 1

10. Diagonalize each of the below matrices. Determine the matrices Q and D such that Q−1 AQ = D is
diagonal. Workout out Q−1 and verify that Q−1 AQ is diagonal and that its diagonal elements are
eigenalues of A. If A is not diagonalizable, state the reason.
(a)  
2 0 0
0 1 1
0 1 1
 (b)  
2 1 −1
 1 4 3 
−1 3 4
(c)  
0 1 0
1 0 0
0 0 0

11. Show that if A is diagonalizable so that Q−1 AQ = D, then

Am = QDm Q−1

Use this result to evaluate A1000 where

 
0 2 2
A = 2 0 2
2 2 0

12. Can a singular matrix be diagonalized? Explain.

13. Use the method of diagonalization to obtain the general solution of the given system of differential
equations, where primes denote d/dt
(a)

x ′ = 4x + y + 3z
y′ = x − z
z′ = 3x − y + 4z

(b)

x ′ = x + 4y
y′ = x + y

14. For a rigid body B , the moment of momentum HO about the origin O and the angular velocity ω
⃗ are
⃗:
related as HO = Iω     
( Hp )x Ixx − Ixy − Ixz ωx
 ( H p )y  = − Iyx Iyy − Iyz   ωy 
( H p )z − Izx − Izy Izz omegaz
where Z Z
Ixx = (y2 + z2 )dm Ixy = xy dm
B B
.. ..
. .
where Ixx , Iyy , Izz are known as the moments of Inertia of the body B about the cartesian axes and
Ixy , Iyz , Ixz are the product of inertia of B , dm is the infinitesimal mass. The above equation is simplest
to deal with if the matrix of inertia is diagonal. This diagonalization is possible if the coordinates are
rotated to a new set of coordinates x̃, ỹ, z̃ so the rotated I, namely Ĩ is diagonal. If x = Qx̃, where
x = [ x y z] T and x̃ = [ x̃ ỹ z̃] T , then HO = Iω
⃗ becomes

QH˜O = IQω̃
⃗

. Thus
H˜O = Q−1 IQω̃
⃗
Here Q−1 IQ = Ĩ is diagonal. Compute the principal interia and principal axes for each of the following
shapes

a)

b)