Sei sulla pagina 1di 10

1 Overview of M147

1. Completeness property of R.
Denition 1. R is an ordered eld containing the N and for every non-empty subset that is bounded
above contains a least upper bound.
S is bounded above means there exists a c R such that x C for all x S. Any such C is an
upper bound. C is a least upper bound for S if C is an upper bound and if D is any upper bound
for S than C D.
People usually called the LUB is also called supremum or sup. The greatest lower bound is called
the inmum or inf.
All properties of R follows from defn.
Decimal representation
Rationals are Dense
Archimedean Property: Given any r Rn N such that n > r.
Continuum Property
2. Sequences
Denition 2. Say (x
n
) converges to a real number L if N N such that |x
n
L| < for all
n N.
Theorem 1.1. (Monotone Convergence) Every bounded monotone sequence converges.
Theorem 1.2. (Bolzano-Weierstrass) Every bounded sequence has a convergent subsequence.
Denition 3. (Cauchy) > 0N N such that |x
n
x
m
| < n, m N.
Theorem 1.3. Every Cauchy sequence converges.
3. Limits + Continuity.
At a A Domain(f) if for all > 0 such that if x A and |x a| < , then |f(x) f(a)| < .
If A = R or A = (c, d) then this denition is equivalent to lim
xa
f(x) = f(a).
Sequential characterization of continuity.
Intermediate Value Theorem
Extreme Value Theorem
4. Dierentiability.
MVT: If f : [a, b] R continuous and is dierentiable on (a, b) then there exists c (a, b) such
that f

(c) =
f(b)f(a)
ba
.
Cauchy MVT.
1
2 Integration Theory
2.1 Denite Integral
Suppose f : [a, b] R is bounded. Partition the interval [a, b] into n subintervals [t
i1
, t
i
] for i [1, n].
For f : [t
i1
, t
i
] R is a bounded function. Let
M
i
= sup {f(t) : t [t
i1
, t
i
]} (1)
m
i
= inf {f(t) : t [t
i1
, t
i
]} (2)
(3)
In particular if f is continuous then M
i
and m
i
is attained.
Denition 4 (Riemann Sums). Let t
i
= t
i
t
i1
.
U(f, P) =
n

i=1
M
i
t
i
(4)
L(f, P) =
n

i=1
m
i
t
i
(5)
We say U(f, P) is the upper riemann sum and L(f, P) is lower Riemann sum.
Theorem 2.1. L(f, P) Area U(f, P).
Denition 5. Let Q be a renement if Q is a partition which includes all the points in the partition P.
Theorem 2.2. If Q is a renment of P then U(f, P) U(f, P) and L(f, P) L(f, Q). Furthermore,
for any two partitions P
1
, P
2
, take Q = P
1
P
2
then L(f, P
1
) L(f, Q) U(f, Q) U(f, P
2
).
Look at {L(f, P) : P partition}. Every U(f, P) is an upper bound for this set. This set is bounded
above and therefore has a lowest upper bound. Let S = sup {L(f, P)}. Thus, S U(f, P). Similarly,
{U(f, P)} is a set bounded below let I = inf {U(f, P)}. Thus, S I.
Now,
L(f, P) S I U(f, Q)
Denition 6. We say f is integrable on [a, b] if sup {L(f, P)} = inf {U(f, P)} and we write
_
b
a
f =
common value. If f 0 then we dene area under f between a and b as
_
b
a
f.
Example 1. f(x) = c in [a, b]. Then U(f, P) = c(b a) = L(f, P) =
_
b
a
f.
Example 2. f(x) =
_
1, x Q
0, x Q
. sup = 1 and inf = 0. Thus not integrable.
Theorem 2.3. Suppose f : [a, b] R is bounded. Then f is integrable over [a, b] if and only if for every
> 0 there is a partition P

such that U(f, P

) L(f, P

) < .
Proof. Assume f is integrable. That means I = S. Fix > 0. Consider S

2
< S. So there must exist
a partition P
1
such that L(f, P
1
) > S

2
. Similarly there is a partition P
2
such that U(f, P
2
) < I +

2
.
Let P

= P
1
P
2
. Then,
U(f, P

) L(f, P

) U(f, P
2
) L(f, P
1
) I +

2
(S

2
) = I S +
Conversely, assume we have such a partition P

. Suppose I = S, let = I S > 0. Then I U(f, P

), S
L(f, P

). Then = U(f, P

)L(f, P

) < but this is a contradiction. Therefore I = S and f is integrable.


za
2
Example 3. f(x) = x
2
on [0, 1].
Choose a uniform partition P
n
: 0t
0
< t
1
=
1
n
< t
2
=
2
n
.
t
i1
=
i 1
n
, t
i
=
i
n
(6)
M
i
= t
2
i
(7)
m
i
= t
2
i1
(8)
U(f, P) =

(
i
n
)
2
(
i
n

i 1
n
) =
1
n
3
n

i
2
=
1
n
3
n(n + 1)(2n + 1)
6
(9)
L(f, P) =

(
i 1
n
)
2
(
i
n

i 1
n
) =
1
n
3
n1

i
2
=
1
n
3
(n 1)n(2(n 1) + 1)
6
(10)
U L =
1
n
(11)
Let > 0. Pick N >
1

. For all n N we have U L =


1
n
< . Thus, f is integrable.
L(f, P), U(f, P),
2
3
. Thus
_
1
0
x
2
=
2
3
.
3 Uniform Continuity
Recall:
Denition 7. f is continuous at x
0
D(f) if > 0 > 0 such that if x D(f) aand |x x
0
| < ,
then |f(x) f(x
0
)| < .
depends on both x
0
and !
Denition 8. f is uniformly continuous if > 0 > 0 such that if x, y D(f) and |x y| < then
|f(x) f(y)| < .
Uniform continuity implies continuity, but continuity uniform cont.
Example 4. f(x) = x
2
is contiuous but not uniformly continuous.
Proof. Let = 1. Say > 0 Take x = a +

2
, y = a.
|x y| =

2
<
|f(x) f(y)| = |(a +

2
)
2
a
2
| = |a
2
+ a +

2
4
a
2
| a
But if a >
1

. This is greater than 1.


Example 5. f(x) =
1
x
on (0, 1). Let > 0. Say works. Take x =
1
N
, y =
2
N
.
Example 6. f : [a, b] R or (a, b), with a bounded derivative, then f is uniformly continuous.
Proof. Let > 0 and =

C
. If x, y D(f) by MVT
|f(x) f(y)| = |f

(t)| |x y| C|x y| < C = C



C
= .
for some t (x, y).
Big theorem:
3
Theorem 3.1. If f : [a, b] R is continuous then f is uniformly continuous.
Proof. (By Contradiction) Suppose this is false. There is a choice of such that no works. In particular,
=
1
n
fails for each n. Hence for each n N there are points x
n
, y
n
[a, b] such that |x
n
y
n
| < but
|f(x) f(y)| .
Consider (x, n), (y
n
) in [a, b].
(x
n
) is bounded sequence. Bolzano-W thm, it has a convergent subsequence (x
n
k
) x
0
[a, b].
Look at (y
n
k
) with the same indices. This has convergent subsequence that converges to y
0
[a, b]. Call
seq (y
nkj
). (x
nkj
) x
0
.
f is continuous at x
0
, y
0
. So f(x
nkj
) f(x
0
and f(y
nkj
) f(y
0
).
|f(x
nkj
f(y
nkj
)|
|f(x
0
) f(y
0
)|
|x
nkj
y
nkj
| <
1
nkj
0 as j . Hence 0 = |y
0
x
0
| x
0
= y
0
.
f(x
0
) = f(y
0
). Contradiction.
Proof. Suppose not. There is some such that no works.In particular, =
1
n
fails for each n. We will
be able to pick x
n
, y
n
D(f) such that |x
n
y
n
| < =
1
n
but |f(x
n
) f(y
n
)| . The sequence (x
n
)
from [a, b] has a convergent subsequence (x
n
k
), say limit x
0
. Consider (y
n
k
). We have |x
n
k
y
n
k
| <
1
n
k
.
By squeeze theorem, (y
n
k
) x
0
.
Now the closure of the interval becomes important when we have f(x
n
k
) f(x
0
) since x
0
[a, b] and f
is continuous. Similarly for f(y
n
k
) f(x
0
). But |f(x
n
k
) f(y
n
k
)| |f(x
0
) f(y
0
)| = 0.
4 Integration Theory Contd
Theorem 4.1. If f : [a, b] R is continuous, then f is integrable.
Proof. Note that f is bounded by the EVT and f is uniformly continuous. We will use the criteria that
f is integrable if for every > 0 there exists a parition P such that U(f, P) L(f, P) < .
Let > 0. Pick from the defn of uniform continuity such that |x y| < so |f(x) f(y)| <

ba
. Take
P such that t
i
t
i1
< .
Assume M
i
= sup f|
[t
i1
,t
i
]
= f(d
i
) for d
i
in[t
i1
, t
i
]. Similarly for m
i
= f(c
i
). These points are attained
because of the EVT.
U(f, P) L(f, P) =

(f(d
i
) f(c
i
))(t
i
t
i1
)
Since c
i
, d
i
[t
i1
, t
i
] thus, |c
i
d
i
| < . By uniform continuity, |f(c
i
) f(d
i
)| <

ba
. So,
U(f, P) L(f, P) <


b a
(t
i
t
i1
) =
Theorem 4.2. Every monotonic function f : [a, b] R.
Proof. Assume f is increasing. f is bounded above by f(b) and below by f(a). Take a uniform partition
P
n
such that a = t
0
< < t
n
= b where t
i
t
i1
=
ba
n
.
U(f, P) =

M
i
(t
i
t
i1
) =

f(t
i
)
b a
n
L(f, P) =

m
i
(t
i
t
i1
) =

f(t
i1
)
b a
n
U(f, P) L(f, P) =

(f(t
i
) f(t
i1
))(
b a
n
) =
b a
n
(f(b) f(a))
Let > 0. Choose n =
(ba)(f(b)f(a))

. Therefore U(f, P) L(f, P) < .


4
Exercise 1. Prove that a monotionic function on [a.b] can have countable many discontinuities.
Theorem 4.3. f : [a, b] R is bounded and continuous except at nitely many points c
1
, . . . , c
s
.
Proof. We begin by surrounding the discontinuities with very small sub intervals. r
0
= a, . . . , r
2s+1
= b
such that c
j
(r
2j1
, r
j
) and

s
j=1
(r
2j
r
2j1
) <

2 sup |f|
.
Look at [r
2j
, r
2j+1
]. Since no c
i
[r
2j
, r
2j+1
] and thus f|
[r
2j
,r
2j+1
]
is uniformly continuous. Pick
j
> 0
from uniform continuouty such that if x, y [r
2j
, r
2j+1
] and |x y| <
j
then |f(x) f(y)| <

2(ba)
.
Now partition each [r
2j
, r
2j+1
] into subintervals of length <
j
. Out partition P will consist of all
r
j
and all the points from the partition of [r
2j
, r
2j+1
] for j = 0, . . . .s. In the following let (1) such that
t
i1
= r
2j1
, t
i
= r
2j
, and (2) such that t
i
s come from a parition of some [r
2j
, r
2j+1
].
U(f, P) L(f, P) =

i
(M
i
m
i
)(t
i
t
i1
) = (1) + (2) (12)
(1) M
i
sup |f| (13)
m
i
sup |f| (14)
(1) 2 sup |f|

(t
i
t
i1
) (15)
2 sup |f|
_

4 sup |f|
_
(16)
=

2
(17)
(2) M
i
= f(c
i
) = max(f|
[r
2j
,r
2j+1
]
) (18)
m
i
= f(d
i
) = min(f|
[r
2j
,r
2j+1
]
) (19)
M
i
m
i
= f(c
i
) f(d
i
) <

2(b a)
(20)
(2)

(M
i
m
i
)(t
i
t
i1
)

2(b a)
(b a) =

2
(21)
Conclusion: U(f, P) L(f, P) = (1) + (2) <

2
+

2
= .
Denition 9. (Measure Zero) We say E R has measure zero if > 0 there are countable many open
intervals I
j
, j = 1, 2, . . . such that E

j=1
I
j
and

j=1
lengthI
j
< .
Example 7. Finite set
Q
Any countable set has measure zero.
E = [0, 1] is NOT measure zero
E is irrationals in [0, 1] are not of measure zero.
Theorem 4.4. (Lebesgue) Bounded function f : [a, b] R is integrable if and only if the set of disconti-
nuities of f is of measure zero.
5 Arc Length Derivation
Denition 10. Arc Length of a continuously dierentiable f : [a, b] R. We mean sup {||f(t
j
) f(t
j1
||}
where ||f(t
j
)f(t
j1
)|| is the length of the line segment joining (t
j1
, f(t
j1
)), (t
j
, f(t
j
)) =
_
(t
j
t
j1
)
2
+ (f(t
j
) f(t
j1
))
2
.
5
f(t
j
) f(t
j1
) = f

(c
j
)(t
j
t
j1
)
for some c
j
(t
j1
, t
j
) by MVT. Thus,
t
j
=
_
(t
j
t
j1
)
2
+ f

(c
j
)
2
(t
j
t
j1
)
2
= (t
j
t
j1
)
_
1 + (f

(c
j
))
2
n

i=1
L
j
=
n

j=1
_
1 + (f

(c
j
))
2
(t
j
t
j1
(22)
Take F(x) =
_
1 + (f

(x))
2
then,
Have observed

f(t
j
) f(t
j1
) = R(F, P):
L(F, P) R(F, P) U(F, P)
If Q is a renement of P then:
R(F, P) =

j
L
(P)
j

i
L
(Q)
i
R(F, Q)
Let P, P

be any parition and Q is a common renement,


R(F, P) R(F, Q) U(f, Q) U(F, P

)
sup
P
L(F, P) sup
P
R(F, P) U(F, P

) P

Therefore,
sup
P
R(F, P) inf
P
U(F, P)
Thus, sup
P
R(F, P) =
_
b
a
F which means that the arclength R(F, P) is
_
b
a
_
1 + (f

(x))
2
.
6 Improper Integrals
Denition 11.
_

a
f(x)dx = lim
s
_
s
a
f(x)dx
_

f =
_
a

f +
_

a
f
Example 8.
_

1
dx
x
3/2
= lim
s
_
s
1
x
3/2
dx = lim
s
(2s
1/2
+ 2) = 2
Exercise 2.
_

a
x
p
dx
converges if p > 1 and diverges if p 1, provided a > 0.
Theorem 6.1. (Comparison Test) Suppose 0 f(x) g(x)x a. If
_

a
g converges, then
_

a
f
converges.
6
Proof. For n N, let n > a. Let
I
n
=
_
n
a
f(x)dx
I
n+1
I
n
=
_
n+1
a
f
_
n
a
f =
_
n+1
n
f 0
(I
n
)

n=n
is an increasing sequence.
I
n

_
n
a
g
_

a
g R
since
_

a
g converges.
So (I
n
) is bounded above, hence it converges by MCT. Still must check lim
s
_
s
a
f converges. Given s,
pick n such that n < s n + 1.
I
n
=
_
n
a
f
_
s
a
f
_
n+1
a
f = I
n+1
By squeeze theorem, since I
n
, I
n+1
L, then
_
s
a
f L.
Example 9.
_

0
e
x
2
dx =

2
Proof it converges: Note: e
x
2
e
x
for x 1.
lim
s
_
s
0
e
x
2
dx.
_
s
0
=
_
1
0
e
x
2
. .
C
+
_
s
1
e
x
2
dx
Using the comparison test since e
x
2
e
x
. Check if
_

1
e
x
converges.
lim
s
_
s
1
e
x
dx = lim
s
e
x

s
1
= lim
x
e
x
+ e
1
= e
1
Example 10. Find the volume of the innitely llong horn generated by rotating y =
1
x
, to the rght of
x = 1 around the x-axis.
V = lim
s

_
s
1
1
x
2
dx = lim
s
x
1

s
1
= lim
s
(
1
s
+ 1) =
What is its surface area?
SA = lim
s
_
s
0
2f(x)
_
1 + f

(x)
2
Denition 12. Say f is unbounded at a. Dene
_
b
a
f = lim
0
+
_
b
a+
f
provided f is integrable over each [a + , b].
7
Example 11.
_
1
1
dx

1 x
2
7 Series
Denition 13.

n=1
a
n
= lim
N
N

n=1
a
n
.
Theorem 7.1. (Geometric Series) S
n
=

N
1
r
n
=
r(1r
n
)
1r
. If |r| < 1 then r
n
0 as N hence
S
n
. Therefore:

1
r
n
=
r
1 r
If |r| 1,

N
1
a
n
diverges.
Example 12. (Telesoping Sum)

1
1
n

1
n(n + 1)
=

1
n(n + 1)
S
n
=
N

1
_
1
n

1
n + 1
_
= 1
1
2
+
1
2

1
3
+
1
3

1
4
+ +
1
N

1
N + 1
= 1
1
N + 1
Therefore S
n
1 as n .
Facts:
1. If

1
a
n
and

1
b
n
both converges then so does

1
a
n
b
n
.
Proof. Write S
(a)
n
=

a
n
, S
(b)
n
=

b
n
. Say S
n
(a) S
(a)
and sim for S
(b)
.
N

1
a
n
+ b
n
= S
(a)
n
+ S
(b)
n
S
(a)
+ S
(b)
2. If

a
n
converges then so does

ca
n
= c

a
n
.
3. If a
n
0 then

a
n
converges if and only if (S
n
)

n=1
is bounded.
Proof. If

a
n
converges then by defn (S
n
) converges (S
n
) bounded.
Suppose (S
n
) bounded. We have S
n+1
S
n
= a
n+1
0 S
n+1
S
n
. Therefore (S
n
) is increasing
. Hence by MCT. (S
n
) converges. So be defn.

a
n
converges.
4. If

a
n
converges then a
n
0.
Proof. Since (S
n
) is convergent, its Cauchy. So given > 0. We can nd N such that |a
n+1
| =
|S
n
S
n+1
| < if n N a
n
0.
8
Example 13. (Harmonic Series)

n=1
1
n
Let us compare with
_

1
dx
x
. (Pictorially, the rectangles all sit above the curve in the discrete sum).
Formally, the discrete sum is the upper riemann sum when taking the partition of evenly spaced out
along the integers.

1
n

_
dx
x
= ln(N + 1)
So hence S
n
is unbounded and hence

1
n
diverges.
Example 14.

1
1
n
1+
for > 0, this sum converges. Compare with
_

1
dx
x
1+
=
x

S
n
= Area of Rect under curve = 1 +
1
2
1+
+ +
1
N
1+

_
N
1
1
x
1+
+ 1 =
x

N
1
+ 1 =
!


1
N

+ 1
1

+ 1
So (S
n
) is bounded and thus

1
n
1+
converges.
Theorem 7.2. (Integral Test) If f is positive and decreasing on [1, ) and f(n) = a
n
, then

a
n
converges if and only if
_

1
f converges.
Proof. Proof by contrapositive, assume
_
f diverges then

a
n
diverges.
Example 15.

1
n
p
converges if and only if p > 1.
Theorem 7.3. (Comparison Test) If 0 a
n
b
n
n n
0
. If

b
n
converges then so does

a
n
.
Proof. Write S
(a)
n
and for b as well. For large N, S
(a)
n
=

n
0
1
a
n
+

N
n
0
a
n
A+S
(b)
n
C for all N since

b
n
converges. So (S
(a)
n
) is bounded and thus

a
n
converges.
Example 16.

1
2n + 1
1
2n+1

1
3n
and

1
3n
diverges.
Example 17.

2n 1
n
3
+ 1

3
n
2
Theorem 7.4. (Ratio Test) If 0 a
n
, b
n
and lim
an
bn
= c = 0. Then

a
n
converges if and only if

b
n
converges.
9
8 Area
Denition 14. If f 0 continuously dierentiable then SA < V ol < .
Proof. (Sketch)
SA = 2
_

a
f(x)
_
1 + f

(x)
2
dx
2
_

a
f(x)
Claim 1. For all nx
n
n such that f(x
n
) 1. If not, then n
0
st f(x) 1x n
0
.
2
_

a
f 2
_

n
0
2(N n
0
)
Proof. (Real) S
n
= solid by rotating f on [a, x
n
] around xaxis.
V (S
n
) Vol of sphere that has the same SA (By Isoperimetric property)
SA(S
n
) = f(a)
2
+ f(x
n
)
2
+ f(x
n
)
2
+ 2
_

a
f(x)
_
1 + f

(x)
2
dx

_
f(a)
2
+ 1
_
+ 2
_

a
f(x)
_
1 + f

(x)
2
dx =
_
f(a)
2
+ 1
_
+ SA = C <
V olS
n
V ol sphere which has surface area = C.
V ol Solid of revn = limV (S
n
) Vol sphere which has surface area = C < .
10

Potrebbero piacerti anche