You are on page 1of 4

0.

1 B and Functions
We dene the B and functions (Eulers functions) and briey state the main properties and applications to
computing some improper integrals.
Denition. The (improper) integral
B(p; q)
def
=
1
_
0
x
p1
(1 x)
q1
dx
is convergent for all p; q > 0 , therefore the " B (beta) function" makes sense.
Denition. The (improper) integral
(p)
def
=
+1
_
0
x
p1
e
x
dx
is convergent for all p > 0 , therefore the " (gama) function" makes sense.
Proof. To prove these improper integrals are convergent is easy.
For B(p; q) we split as ( for any c 2 (0; 1) )
B(p; q) =
1
_
0
x
p1
(1 x)
q1
dx =
c
_
0
x
p1
(1 x)
q1
d
. .
x
A
+
1
_
c
x
p1
(1 x)
q1
dx
. .
B
The improper integral "A" is "improper" only at x = 0 , "B" is "improper" only at x = 1
Since for p 1 and q 1 we have
lim
x&0
x
p1
(1 x)
q1
= 0 , lim
x%1
x
p1
(1 x)
q1
= 0
A and B are improper only for 0 < p < 1 , 0 < q < 1 , since in this case
lim
x&0
x
p1
(1 x)
q1
= lim
x&0
1
x
1p
(1 x)
q1
=
1
+0
= +1
lim
x%1
x
p1
(1 x)
q1
= lim
x%1
x
p1
1
(1 x)
q1
=
1
+0
= +1
Finally for p < 1 , q < 1 , we use the limit comparison test.
We compare A to
c
_
0
1
x
1p
dx and B to
1
_
c
1
(1 x)
1q
dx
improper integrals which are convergent since 0 < 1 p < 1 and 0 < 1 q < 1
Next compute the limits
lim
x&0
x
p1
(1 x)
q1
1
x
1p
= lim
x&0
(1 x)
q1
= 1 , lim
x%1
x
p1
(1 x)
q1
1
(1x)
1q
= lim
x%1
x
p1
= 1
Consequently, by limit comparison test the improper integrals are convergent, which implies the integral dening
B(p; q) is also convergent.
For (p) we split as
(p) =
+1
_
0
x
p1
e
x
dx =
1
_
0
x
p1
e
x
dx
. .
C
+
+1
_
1
x
p1
e
x
dx
. .
D
1
The integral C is "improper" only at x = 0 and only for p < 1 , since
lim
x&0
x
p1
e
x
= 0 for all p > 1 , lim
x&0
x
p1
e
x
= 1 for p = 1
lim
x&0
x
p1
e
x
= lim
x&0
1
x
1p
e
x
=
1
+0
= +1 for all p < 1
The integral D is "improper" only at +1 .
We use limit comparison test for p < 1
we compare C to
1
_
0
1
x
1p
dx , which is convergent since 1 p < 1
lim
x&0
x
p1
e
x
1
x
1p
= lim
x&0
e
x
= 1
consequently, by limit comparison test, the integral C is convergent.
we compare D to
+1
_
1
e
x=2
dx , which is convergent
lim
x!+1
x
p1
e
x
e
x=2
= lim
x!+1
x
p1
e
x=2
= 0 for all p > 0
consequently, by limit comparison test, the integral C is convergent.
Finally C and D convergent ) the integral dening (p) is convergent for all p > 0

Properties.
1.
(1) = 1
2. The following property suggests that may be considered a "generalization" of the "factorial" function
(which makes sense only for natural numbers).
(p) = (p 1) (p 1) for all p > 1
(n) = (n 1)! for all n 2 N
3.
B(q; p) = B(p; q) for all p; q > 0
4.
B(
1
2
;
1
2
) =
5.
B(p; q) =
1
_
0
x
p1
(1 x)
q1
dx = 2
=2
_
0
(sint)
2p1
(cos t)
2q1
dt or conversely
=2
_
0
(sint)
m
(cos t)
n
dt =
1
2
B(
m + 1
2
;
n + 1
2
)
6.
B(p; q) =
1
_
0
x
p1
(1 x)
q1
dx =
+1
_
0
t
p1
(t + 1)
p+q
dt
2
7.
B(p; q) =
(p) (q)
(p + q)
for all p; q > 0
8.
(p)(q) =

sinp
=

sinq
for all p + q = 1 , p; q > 0
Proof.
1.
(1) =
+1
_
0
e
x
dx = e
x

+1
0
= lim
x!+1
(e
x
) (e
0
) = 1 lim
x!+1
1
e
x
= 1
2.
(p) =
+1
_
0
x
p1
e
x
dx = by parts = x
p1
__
e
x
_
. .
e
x

+1
0

+1
_
0
_
x
p1
_
0
__
e
x
_
dx =
=
_
lim
x!+1
x
p1
e
x

0
p1
e
0
_

+1
_
0
(p 1)x
p2
(e
x
)dx = 0 0 + (p 1)
+1
_
0
x
p2
e
x
dx = (p 1) (p 1)
For natural numbers n we get
(n) = (n 1)(n 1) = (n 1)(n 2)(n 2) = :::
::: = (n 1)(n 2):::3 2 1 (1) = (n 1)!(1)
3.
B(q; p) =
1
_
0
x
q1
(1 x)
p1
dx = the change of variable x = 1 t leads to
=
0
_
1
(1 t)
q1
t
p1
(dt) =
1
_
0
(1 t)
q1
t
p1
dt = Bp; q)
4. We use the change of variable x = sin
2
t , dx = 2 sint cos tdt we have x &0 )t &0 si x %1 )t %

2
,
we get
B(
1
2
;
1
2
) =
1
_
0
1
_
x(1 x)
dx =
=2
_
0
1
_
sin
2
t(1 sin
2
t)
2 sint cos tdt =
=2
_
0
2 sint cos t
sint cos t
dt =
=2
_
0
2dt =
5. Again we use the change of variable x = sin
2
t , dx = 2 sint cos tdt , just as before, we get
B(p; q) =
1
_
0
x
p1
(1 x)
q1
dx ==
=2
_
0
(sin
2
t)
p1
(1 sin
2
t)
q1
2 sint cos tdt = 2
=2
_
0
(sint)
2p1
(cos t)
2q1
dt
6. We use the change of variable x =
t
t+1
, and clearly x &0 ) t &0 , x %1 ) t !
1
+0
= +1
dx =
_
t
t + 1
_
0
dt =
t + 1 t
(t + 1)
2
dt =
1
(t + 1)
2
dt
B(p; q) =
1
_
0
x
p1
(1 x)
q1
dx ==
+1
_
0
_
t
t + 1
_
p1
_
1
t
t + 1
_
q1
1
(t + 1)
2
dt =
+1
_
0
t
p1
(t + 1)
p+q
dt
3
7. 8. We omit the proof, it is not elementary and irrelevant for the purpose of this presentation.
Here are a few applications of these properties, to computing some denite or improper integrals .
Exemples.
1) Prove that
(
1
2
) =
p

Solution.
= B(
1
2
;
1
2
) =
(
1
2
) (
1
2
)
(
1
2
+
1
2
)
=
(
1
2
)
2
(1)
) (
1
2
) =
p

2) Compute the integral


=2
_
0
(sint)
6
(cos t)
8
dt
Solution.
=2
_
0
(sint)
6
(cos t)
8
dt =
1
2
B(
6 + 1
2
;
8 + 1
2
) =
1
2
(
7
2
) (
9
2
)
(
7
2
+
9
2
)
=
First we have
(
7
2
+
9
2
) = (8) = 7!
(
9
2
) =
7
2
(
7
2
) =
7
2
5
2
(
5
2
) =
7
2
5
2
3
2
(
3
2
) =
7
2
5
2
3
2
1
2
(
1
2
)
(
7
2
) =
5
2
(
5
2
) =
5
2
3
2
(
3
2
) =
5
2
3
2
1
2
(
1
2
)
Then we nalise
=2
_
0
(sint)
6
(cos t)
8
dt =
1
2
(
7
2
) (
9
2
)
(8)
=
5
2
3
2
1
2
(
1
2
)
7
2
5
2
3
2
1
2
(
1
2
)
7!
3) Compute the improper integral
+1
_
0
1
(x
3
+ 1)
3
dx
Solution.
We use the change of variable x
3
= t , x !0 ) t !0 , x !+1 ) t !+1 , x = t
1=3
, dx =
1
3
t
1=31
dt
, we get
+1
_
0
1
(x
3
+ 1)
3
dx =
+1
_
0
t
1=31
(t + 1)
3
dt = B(
1
3
; 3
1
3
) = B(
1
3
;
8
3
) =
(
1
3
) (
8
3
)
(
1
3
+
8
3
)
=
=
(
1
3
)
5
3
(
5
3
)
(3)
=
(
1
3
)
5
3
2
3
(
5
3
)
2!
=
5
3
1
3
(
1
3
)(
5
3
) =
5
9

sin

3
=
5 2
9
p
3

4) Compute the value of Poisson integral


+1
_
0
e
x
2
dx
Solution.
We use the change of variable x
2
= t ) x = t
1=2
, dx =
1
2
t
1=21
dt , x ! 0 ) t ! 0 , x ! +1 )
t !+1 , we get
+1
_
0
e
x
2
dx =
+1
_
0
e
t
1
2
t
1
2
1
dt =
1
2
(
1
2
) =
p

2
4

You might also like