HW Apr 5 Sols

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Math 432 - Real Analysis II

Solutions to Homework due April 5


Question 1. The standard dot product on Rn is given by
hx, yi = xT y.
In this question, we will use positive-definite matrices to create new dot products on Rn .
(a) Consider the 3 × 3 matrix  
2 −1 0
M =  −1 2 −1  .
0 −1 2
We will define a new dot product on R3 via the rule
hx, yi = xT M y.
Write x = (x1 , x2 , x3 ). Show that hx, xi ≥ 0 by writing hx, xi as a sum of squares.
(b) Use (a) to show that our new dot product hx, yi = xT M y is positive-definite.
(c) In general, a matrix M is called positive-definite if xT M x ≥ 0 for all x ∈ Rn and xT M x = 0 only for
x = 0. Show that if M is a positive-definite, symmetric matrix, then hx, yi = xT M y gives an inner
product on Rn by checking that it satisfies the three inner product axioms.

Solution 1.
(a) Letting x = (x1 , x2 , x3 ), we get
  
2 −1 0 x1
xT M x = (x1 , x2 , x3 )  −1 2 −1   x2  =
0 −1 2 x3
 
x1
(2x1 − x2 , −x1 + 2x2 − x3 , −x2 + 2x3 )  x2  =
x3
2x21 − x1 x2 − x1 x2 + 2x22 − x2 x3 − x2 x3 + 2x23 =
x21 + (x21 − 2x1 x2 + x22 ) + (x22 − 2x2 x3 + x23 ) + x23 =
x21 + (x1 − x2 )2 + (x2 − x3 )2 + x23 ≥ 0.

(b) From (a), we know that hx, xi ≥ 0. Clearly, h0, 0i = 0. Conversely, if hx, xi ≥ 0, then x21 + (x1 − x2 )2 +
(x2 − x3 )2 + x23 = 0. The only way for this to occur is if x1 = 0, x1 = x2 , x2 = x3 , and x3 = 0. In other
words, x = 0.
(c) Since M is positive-definite, then by definition, hx, xi = xT M x ≥ 0 with equality if and only if x = 0.
Since M is symmetric, then M T = M . Thus,
hx, yi = xT M y = xT M T y = (yT M x)T .
Since this last term is simply a number (a 1 × 1 matrix), it is equal to its own transpose and we have
that this is equal to
yT M x = hy, xi.
Thus, symmetry holds. For linearity, note that
hαx, yi = (αx)T M y = α(xT M y) = αhx, yi,
and
hx + y, zi = (x + y)T M z = (xT + yT )M z = xT M z + yT M z = hx, zi + hy, zi.
Thus, bilinearity holds.

1
Question 2. In class, we learned how to define the norm in an inner product space in terms of its dot
product. In this question, we will learn an equation that gives the dot product in terms of the norm.

(a) Let V be a real inner product space and x, y ∈ V . Show that


1
||x + y||2 − ||x − y||2 .

hx, yi =
4

(b) Let V be a complex inner product space and x, y ∈ V . Show that


1
||x + y||2 − ||x − y||2 + ||x + iy||2 i − ||x − iy||2 i

hx, yi =
4

Solution 2.

(a) Notice that


||x + y|| = hx + y, x + yi = hx, xi + hx, yi + hy, xi + hy, yi.
Since V is a real vector space, then the two middle terms are equal and we have that the above gives

hx, xi + 2hx, yi + hy, yi.

Similarly, we have that


||x − y||2 = hx, xi − 2hx, yi + hy, yi.
Subtracting these two terms, we get 4hx, yi, which is equivalent to our desired statement.

Question 3. In this question, we will show that the converse to the Pythagorean Theorem holds in real
inner product spaces.

(a) Let z = a + bi ∈ C be a complex number. Show that z + z = 2Re(z), where Re(z) = a is the real part
of z.

(b) Use (a) to prove that in a (real or complex) inner product space,

hx, yi + hy, xi = 2Re(hx, yi).

(c) Use (b) to prove that if V is a real inner product space and if

||x + y||2 = ||x||2 + ||y||2 ,

then x and y are orthogonal.

Solution 3.

(a) If z = a + bi, then z = a − bi and we get

z + z = (a + bi) + (a − bi) = 2a = 2Re(z).

2
(b) By conjugate symmetry, we have that
hy, xi = hx, yi.
Thus,
hx, yi + hy, xi = hx, yi + hx, yi = 2Re(hx, yi).
In particular, if V is a real vector space, then hx, yi is a real number and thus equal to its own real part.
So, in this case
hx, yi + hy, xi = 2hx, yi.

(c) Notice that


||x + y||2 = hx, xi + hx, yi + hy, xi + hy, yi = ||x||2 + 2hx, yi + ||y||2 .
If this is equal to ||x||2 + ||y||2 , then we are left with 2hx, yi = 0, leaving us with hx, yi = 0, giving us
that x and y are orthogonal.

Question 4. In class, we defined the p-norm on Rn for p ≥ 1 to be given by

n
!1/p
X
p
||x||p = |xk | .
k=1

(a) Use the Cauchy-Schwarz Inequality to show that for any vector ||x||,

||x||1 ≤ n||x||2 .

(b) Show that this inequality is sharp. To do this, for every n, find a specific vector x ∈ Rn such that

||x||1 = n||x||2 .

Solution 4.

(a) Beginning with the left-hand side, we have that


n
X n
X
||x||1 = |xk | = |xk | · 1 = h(|x1 |, |x2 |, . . . , |xn |), (1, 1, . . . , 1)i ≤
k=1 k=1

n
!1/2 n
!1/2
X
2
X √
|xk | 1 = n||x||2 .
k=1 k=1

(b) Consider the vector x = (1, 1, . . . , 1). Notice that ||x||1 = n and that ||x||2 = n. Thus,
√ √ √
||x||1 = n = n n = n||x||2 .

Question 5. In this question, we will solidify the relationship between normed vector spaces and metric
spaces. In particular, we will show that every normed vector space is a metric space. To do so, let V be a
(real or complex) normed vector space. Consider the metric d(·, ·) on V given by

d(x, y) = ||x − y||.

Show that d is indeed a metric on V by showing that it satisfies the three metric axioms.

Solution 5.

3
First we show that d is positive definite. Note that by definition of the norm, d(x, y) ≥ 0. Furthermore,
if x = y, we have that d(x, y) = ||0|| = 0. Conversely, if d(x, y) = 0, then

||x − y|| = 0.

By properties of norms, we have that x − y = 0 and thus x = y. So, d is positive definite.


To show that d is symmetric, note that

(.x, y) = ||x − y|| = || − 1(y − x)|| = | − 1| ||y − x|| = d(y, x).

Finally, to show that d satisfies the triangle inequality, we have that

d(x, z) = ||x − z|| = ||x − y + y − z|| ≤ ||x − y|| + ||y − z|| = d(x, y) + d(y, z).

Thus, d satisfies the triangle inequality.

You might also like