Exercise Sheet 5
Exercise Sheet 5
Show that the function φ defined as in (1.1) is increasing on the open interval (c, ∞) ∩ I.
(b) Let f be as in (3a). Show that, for any a, b ∈ I with a < b, one has the following:
f (b) − f (a)
∀x ∈ [a, b], f (x) ≤ f (a) + (x − a). (1.2)
b−a
What does that mean geometrically? (Hint: Observe that, for x ∈ (a, b],
f (b) − f (a) f (x) − f (a) f (b) − f (a)
f (x) ≤ f (a) + (x − a) ⇐⇒ ≤ .
b−a x−a b−a
Does this not suggest us to apply (3a) with c = a?)
(4) Let I ⊆ R be an open interval, a ∈ I, f, g : I −→ R. Assume that both f (n+1) and g (n+1)
exist. Then for any x ∈ I \ {a}, there exists c in the interior of the interval with x and a as
the end points such that the following holds:
n
! n
!
(k) (k)
X f (a) X g (a)
f (x) − (x − a)k g (n+1) (c) = g(x) − (x − a)k f (n+1) (c). (1.3)
k=0
k! k=0
k!
def
Does the special case g(x) = (x − a)n+1 , for all x ∈ I, remind you of something familiar?
(Hint: For fixed x ∈ I \ {a}, consider the following functions:
∞
X ∞
X ∞
X
an = pn − qn .
n=1 n=1 n=1
an+1
(8) Let {an }∞
n=1 be a sequence in R \ {0}. Assume that
−−−→ ℓ, where ℓ ∈ R. Show the
an n→∞
following:
∞
X
(a) ℓ < 1 =⇒ an is absolutely convergent.
n=1
X∞
(b) ℓ > 1 =⇒ an is divergent.
n=1
(c) What happens if ℓ = 1?
2. Additional exercises
(1) Let I and a be as in (1a). Assume that f is strictly increasing and f (a) = 0, where a ∈ I.
For any b ∈ I \ {a}, we define the following sequence recursively:
f (xn )
x1 = b, xn+1 = xn − ′ , ∀n ∈ N. (2.1)
f (xn )
Show that the sequence defined in (2.1) converges to a. Can you connect this to Exercise 5 of
[Exercise Sheet 2, §2]?
(2) Let I and f be as in (3) of §1. Suppose that, for any a, b ∈ I with a < b, the following holds:
f (b) − f (a)
∀x ∈ [a, b], f (x) ≤ f (a) + (x − a).
b−a
Show that f is convex.
(3)∗ Show that given any x ∈ R, there exist x0 ∈ Z and a sequence {xn }∞ n=1 with all xn ∈
∞ ∞
X xn X xn
{0, 1, . . . , 9} such that x = x0 + n
. In this case, we say that x0 + is a decimal
n=1
10 n=1
10n
representation of x and denote that by x0 · x1 x2 . . . Is the decimal representation of a real
number necessarily unique?
m def m def m(m − 1) . . . (m − k + 1)
(4)∗ For m ∈ R and k ∈ N, let = 1 and = . Show that, for
0 k k!
all x ∈ (0, 1), one has
∞
m
X m k
(1 + x) = x .
k=0
k
The aim of the following exercises (5) and (6) is to show one important application of the basic
properties of convex functions that we have developed so far. We will see that the well known AM-
GM inequality can be established using (1a) of §1. Let us first recall the statement of AM-GM
inequality:
AM-GM inequality. Let n ∈ N and x1 , . . . , xn be nonnegative real numbers. Then
x1 + · · · + xn 1
≥ (x1 · · · xn ) n , (2.2)
n
and furthermore, equality occurs in (2.2) if and only if all xi ’s are equal.
Note that the above mentioned inequality is trivial if at least one xi = 0. Hence we assume from
now on that xi > 0, for all i = 1, . . . , n.
(5)∗ For n ∈ N and x1 , . . . , xn > 0, consider the following function:
def x1 + · · · + xn + t 1
f : (0, ∞) −→ R, f (t) = − (x1 · · · xn t) n+1 . (2.3)
n+1
(a) Show that the function defined above in (2.3) is convex.
def 1
(b) Show that t0 = (x1 · · · xn ) n is the unique point of minimum of f .
(c) Calculate f (t0 ).
(6)∗ Prove AM-GM inequality by induction.