Fall 2019 Math 104 - HW 2 Solutions

(If you find any errors, please let your instructor know ASAP)

Problem 1. Let x1 , . . . , xk ∈ Rn be non-zero mutually orthogonal vectors. Show that {x1 , . . . , xk }

must be a linearly independent set.

Solution. That the {xi } are mutually orthogonal and non-zero means that

hxi , xj i = xTi xj = 0, for i 6= j, and hxi , xi i =

6 0.

Suppose that we have a relationship

a1 x1 + a2 x2 + · · · + ak xk = 0.

Then, taking the inner product of both sides with xi , we have that

hxi , a1 x1 i + · · · + hxi , ai xi i + · · · + hxk , ak xk i = hxi , 0i = 0.

Using mutual orthogonality, we see that 0 = hxi , ai xi i = ai hxi , xi i. Because hxi , xi i 6= 0, we see
that ai = 0. As i was arbitrary, ai = 0 for each i. Thus {x1 , . . . , xk } is a linearly independent set.
Problem 2. Let v1 , . . . , vn be orthonormal vectors in Rn . Show that Av1 , . . . , Avn are also or-
thonormal if and only if A is an orthonormal matrix.

Solution. First, suppose that A is orthogonal. Then AT A = AAT = I. We check that the {Avi }
are orthonormal:

hAvi , Avj i = (Avi )T Avj = viT AT Avj = viT (I)vj = viT vj = hvi , vj i.

Thus if i = j, then hAvi , Avj i = hvi , vi i = 1. If i 6= j, then hAvi , Avj i = hvi , vj i = 0. Thus the
{Avi } are orthogonal.
Now suppose that the vectors {Avi } are orthonormal. By the previous problem, the {vi } are
linearly independent, and there are n of them in Rn , so they form a basis. Write A is the basis given
by the {vi }. Then the ith of A is the vectors Avi expressed in the basis given by {vi }. As the {Avi }
are orthonormal, this implies that the columns of A are orthonormal, so A is orthogonal.
Problem 3. Let Pn denote the vector space of polynomials of degree less than or equal to n, and
of the form p(x) = p0 + p1 x + · · · + pn xn , where the coefficients pi are all real. Let PE denote
the subspace of all even polynomials in Pn , i.e., those that satisfy the property that p(−x) = p(x).
Similarly, let PO denote the subspace of all odd polynomials, i.e., those satisfying p(−x) = −p(x).
Show that Pn = PE ⊕ PO .

Solution. By definition, Pn = PE ⊕ PO if PE ∩ PO = 0 and Pn = PE + PO . First, we check that

PE ∩ PO = 0.
Let f ∈ PE ∩ PO . Then f is a polynomial that is both even and odd, i.e., f (−x) = f (x) and
f (−x) = −f (x). Thus f (x) = −f (x), or 2f (x) = 0. Thus f = 0.
Now, we check that Pn = PE + PO . Let f ∈ Pn . Let fE (x) = f (x)+f 2
(−x)
, and let fO (x) =
f (x)−f (−x)
2
. Clearly fE (x) + fO (x) = f (x). Then we claim that fE ∈ PE , and fO ∈ PO . Indeed,

f (−x) + f (− − x) f (x) + f (−x)

fE (−x) = = = fE (x),
2 2
f (−x) − f (− − x) f (x) − f (−x)
fO (−x) = =− = −fO (x).
2 2
Thus we have written each element of Pn as a sum of an element of PE and an element of PO , so
PE + PO = Pn .
Problem 4. This problem examines nilpotent matrices.

(a) Find a square matrix A, whose entries are not all zeros, such that A2 = 0.

(b) Exhibit a nonzero vector that belongs to the nullspace of the matrix that you just constructed.

(c) In general, prove that if a matrix B satisfies B 2 = 0, then it cannot be invertible.

! !
0 1 0 0
Solution. (a) Let A = . Then we see that A2 = .
0 0 0 0
! !
1 0
(b) Let e1 = . Then Ae1 = , so e1 is in the nullspace of A.
0 0

(c) Let B be a matrix such that B 2 = 0. Then we see that

0 = det(0) = det(B · B) = det(B) det(B).

As the only number whose square is 0 is 0, we see that det(B) = 0 and B is not invertible.
Alternatively, suppose that B is invertible and AB is the identity. Then

I = AB = A(I)B = A(AB)B = A2 B 2 = A2 0 = 0,

but the identity matrix is not 0, so B must not have been invertible.
 !
2 3 4
Problem 5. (a) Let A = and consider A as a linear transformation mapping R3 to
8 5 1
R2 . Find the matrix representation with of A with respect to the bases
      
 1
 0 1 
1 , 1 , 0 .
     
 0

1 1 


of R3 and  ! !
3 2
 
,
 1 1 

of R2 .

(b) Programming question.

Solution. (a) We need to express Av for v in the first basis in terms of the second basis. We
compute:  
1 ! ! !
5 3 2
A 1 = = (−21) + 34 ,
 
13 1 1
0
 
0 ! ! !
7 3 2
A 1 = = (−5) + 11 ,
 
6 1 1
1
 
1 ! ! !
6 3 2
A 0 = = (−12) + 21 .
 
9 1 1
1
Therefore, we have that !
−21 −5 −12
A= .
34 11 21

(b) The programming part can be found in another file.