Gamma and Beta Function

Gamma Function: The convergent improper integral ∫ 𝑒 𝑥 𝑑𝑥 , 𝑛 > 0 is

known as second Eulerian integral is called gamma function of 𝑛 and expressed as

(i) 𝛤𝑛 = ∫ 𝑒 𝑥 𝑑𝑥 , 𝑛 > 0.

Another form of Gamma function:

(ii) 𝛤𝑛 = 𝑎 ∫ 𝑒 𝑥 𝑑𝑥 , ( 𝑝𝑢𝑡 𝑥 = 𝑎𝑦 𝑖𝑛 (𝑖))

(iii) 𝛤𝑛 = 2 ∫ 𝑒 𝑥 𝑑𝑥 , ( 𝑝𝑢𝑡 𝑥 = 𝑦 𝑖𝑛 (𝑖))

(iv) 𝛤𝑛 = (𝑚 + 1) ∫ 𝑥 log 𝑑𝑥 , ( 𝑝𝑢𝑡 𝑥 =

−(𝑚 + 1) 𝑙𝑜𝑔𝑦 𝑖𝑛 (𝑖)).

Properties of Gamma function:

(i) 𝛤1 = 1.
(ii) 𝛤(𝑛 + 1) = 𝑛𝛤𝑛.
(iii) 𝛤(𝑛 + 1) = 𝑛!
(iv) 𝛤𝑛 𝛤(1 − 𝑛) = , 0 < 𝑛 < 1.
(v) 𝛤 = √𝜋.
(vi) 2 𝛤𝑛 𝛤(𝑛 + ) = 𝛤(2𝑛)√𝜋.

Proof: 𝛤1 = 1.

we know 𝛤𝑛 = ∫ 𝑒 𝑥 𝑑𝑥 (1)

putting 𝑛 = 1 in (1) we get

𝛤1 = 𝑒 𝑑𝑥

=lim → ∫ 𝑒 𝑑𝑥

= lim[−𝑒 ]
→

∴ 𝛤1 = lim(1 − 𝑒 ) = 1.
→

Proof: 𝛤(𝑛 + 1) = 𝑛𝛤𝑛.

we know 𝛤𝑛 = ∫ 𝑒 𝑥 𝑑𝑥 (2)
 Replacing 𝑛 by (𝑛 + 1) in (2) we get

𝛤(𝑛 + 1) = 𝑒 𝑥 𝑑𝑥

=lim → ∫ 𝑒 𝑥 𝑑𝑥

= lim {[−𝑒 𝑥 ] +𝑛 𝑒 𝑥 𝑑𝑥}

→

= lim { −𝑒 𝑏 }+𝑛 𝑒 𝑥 𝑑𝑥
→

𝑏
= lim { − } + 𝑛𝛤𝑛
→ 𝑒
!
= lim { → − } + 𝑛𝛤𝑛 [Using L’Hospital’s rule n times]

= 0 + 𝑛𝛤𝑛

∴ 𝛤(𝑛 + 1) = 𝑛𝛤𝑛.

Proof: 𝛤(𝑛 + 1) = 𝑛!

𝛤(𝑛 + 1) = 𝑛𝛤𝑛

= 𝑛𝛤((𝑛 − 1) + 1)

= 𝑛(𝑛 − 1)𝛤(𝑛 − 1)

= 𝑛(𝑛 − 1)𝛤((𝑛 − 2) + 1)

= 𝑛(𝑛 − 1)(𝑛 − 2)𝛤(𝑛 − 2)

= 𝑛(𝑛 − 1)(𝑛 − 2)𝛤(𝑛 − 2) … … … .3.2.1𝛤1, if n is positive integer

= 𝑛(𝑛 − 1)(𝑛 − 2)𝛤(𝑛 − 2) … … … .3.2.1.1 (𝛤1 = 1)

: 𝛤(𝑛 + 1) = 𝑛!

Proof: 𝛤 = √𝜋.

𝛤𝑛 = 2 𝑒 𝑥 𝑑𝑥
 1
𝛤( ) = 2 𝑒 𝑑𝑥
2
Let 𝐼 = ∫ 𝑒 𝑑𝑥 = ∫ 𝑒 𝑑𝑦
∴𝐼 =∫ ∫ 𝑒 ( )
𝑑𝑥𝑑𝑦 (1)

Let 𝑥 = 𝑟𝑐𝑜𝑠𝜃, 𝑦 = 𝑟𝑠𝑖𝑛𝜃. Therefore 𝑑𝑥𝑑𝑦 = 𝑟𝑑𝑟𝑑𝜃. As 𝑥 and 𝑦 varies from 0 to

∞ then 𝑟 varies from 0 to ∞ and 𝜃 varies from 0 to 𝜋/2. So (1) becomes
/ ( )
𝐼 =∫ ∫ 𝑒 𝑟 𝑑𝑟 𝑑𝜃
/
𝑜𝑟 𝐼 = ∫ − 𝑒 𝑑𝜃
/
𝑜𝑟 𝐼 = ∫ 𝑑𝜃 =
√
𝑜𝑟 𝐼 =
1
∴ 𝛤( ) = √𝜋
2
( )
we know 𝛤𝑛 =

1 𝛤(−1/2 + 1) 𝛤(1/2)
∴ 𝛤(− ) = = = −2√𝜋
2 −1/2 −1/2
1
3 𝛤(−3/2 + 1) 2𝛤 −
∴ 𝛤(− ) = =− 2 =4 𝜋
√
2 −3/2 3 3
( )
Ex1: Show that 𝛤(𝑛 + ) = √𝜋
( )

1 (2𝑛 + 1) (2𝑛 − 1)
𝛤 𝑛+ =𝛤 =𝛤 +1
2 2 2

2𝑛 − 1 (2𝑛 − 1)
=( )𝛤
2 2

2𝑛 − 1 (2𝑛 − 3)
=( )𝛤 +1
2 2

2𝑛 − 1 2𝑛 − 3 (2𝑛 − 3)
=( )( )𝛤
2 2 2

2𝑛 − 1 2𝑛 − 3 2𝑛 − 5 5 3 1 1
= ………. . . 𝛤
2 2 2 2 2 2 2
 (2𝑛 − 1)(2𝑛 − 3)(2𝑛 − 5) … . .5.3.1
= √𝜋
2
2𝑛 (2𝑛 − 1)(2𝑛 − 2)(2𝑛 − 3)(2𝑛 − 4)(2𝑛 − 5) … . .5.4.3.2.1
= √𝜋
2 2 𝑛. (𝑛 − 1)(𝑛 − 2) … … . .2.1
( )! ( )
∴ 𝛤(𝑛 + ) =
!
√𝜋 = ( )
√𝜋 .

√
Ex:2 Show that ∫ 𝑒 𝑑𝑥 =

𝐿. 𝐻. 𝑆 = ∫ 𝑒 𝑑𝑥, [put 𝑥 = 𝑧 𝑜𝑟 𝑑𝑥 = 𝑧 /
𝑑𝑧]

1 /
= 𝑒 𝑧 𝑑𝑧
2

1
= 𝑒 𝑧 𝑑𝑧
2

1 1 √𝜋
= 𝛤 = .
2 2 2

Ex3: Show that ∫ 𝑒 𝑥 𝑑𝑥 × ∫ 𝑒 𝑑𝑥 = .

√

𝐿. 𝐻. 𝑆 = ∫ 𝑒 𝑥 𝑑𝑥 × ∫ 𝑒 𝑑𝑥, [put 𝑥 = 𝑧 𝑜𝑟 𝑥 = 𝑧 /
𝑜𝑟 𝑑𝑥 = 𝑧 /
𝑑𝑧]

1 / /
1 /
= 𝑒 𝑧 𝑧 𝑑𝑧 × 𝑒 𝑧 𝑑𝑧
4 4

1 / /
= 𝑒 𝑧 𝑑𝑧 × 𝑒 𝑧 𝑑𝑧
16

1
= 𝑒 𝑧 𝑑𝑧 × 𝑒 𝑧 𝑑𝑧
16

1 3 1 1 1 1
= 𝛤( )𝛤( ) = 𝛤( )𝛤(1 − )
16 4 4 16 4 4

= , 𝛤𝑛𝛤(1 − 𝑛) = ,0 < 𝑛 < 1

𝜋 1 𝜋
= = = 𝑅𝐻𝑆
16 1/√2 8√2
 √
Ex4: Show that ∫ 𝑒 𝑥 𝑑𝑥 =

[put 𝑥 = 𝑧 𝑜𝑟 𝑥 = 𝑧 /
𝑜𝑟 𝑑𝑥 = 𝑧 /
𝑑𝑧]

√
Ex5: Show that ∫ 𝑒 𝑥 𝑑𝑥 =

[put 𝑥 = 𝑧 𝑜𝑟 𝑥 = 𝑧 /
𝑜𝑟 𝑑𝑥 = 𝑧 /
𝑑𝑧]
( )
( ) ( )
Ex6: Show that ∫ 𝑥 (1 − 𝑥 ) 𝑑𝑥 =
( )

[put 𝑥 = 𝑧 𝑜𝑟 𝑥 = 𝑧 /
𝑜𝑟 𝑑𝑥 = 𝑧 ( )/
𝑑𝑧]

Ex7: Show that ∫ 𝑑𝑥 = . , 0 < 𝑛 < 1.

/

we know 𝛤𝑛 = ∫ 𝑒 𝑥 𝑑𝑥 (1)

𝛤𝑛
𝑜𝑟, = 𝑒 𝑦 𝑑𝑦, [𝑃𝑢𝑡 𝑥 = 𝑎𝑦𝑖𝑛 (𝑖)]
𝑎

1 1
𝑜𝑟, = 𝑒 𝑦 𝑑𝑦
𝑎 𝛤𝑛

1 1
𝑜𝑟, = 𝑒 𝑦 𝑑𝑦 , [𝑅𝑒𝑝𝑙𝑎𝑐𝑖𝑛𝑔 𝑎 𝑏𝑦 𝑥]
𝑥 𝛤𝑛

𝑐𝑜𝑠𝑏𝑥 𝑐𝑜𝑠𝑏𝑥
𝑜𝑟, 𝑑𝑥 = ( 𝑒 𝑦 𝑑𝑦)𝑑𝑥
𝑥 𝛤𝑛

𝑐𝑜𝑠𝑏𝑥 1
𝑜𝑟, 𝑑𝑥 = 𝑐𝑜𝑠𝑏𝑥 𝑒 𝑦 𝑑𝑦𝑑𝑥
𝑥 𝛤𝑛

𝑐𝑜𝑠𝑏𝑥 1
𝑜𝑟, 𝑑𝑥 = 𝑐𝑜𝑠𝑏𝑥 𝑒 𝑦 𝑑𝑥𝑑𝑦
𝑥 𝛤𝑛

𝑐𝑜𝑠𝑏𝑥 1
𝑜𝑟, 𝑑𝑥 = 𝑦 { 𝑐𝑜𝑠𝑏𝑥 𝑒 𝑑𝑥}𝑑𝑦
𝑥 𝛤𝑛

𝑐𝑜𝑠𝑏𝑥 1 𝑒 (−𝑦𝑐𝑜𝑠𝑏𝑥 + 𝑏𝑠𝑖𝑛𝑏𝑥

𝑜𝑟, 𝑑𝑥 = 𝑦 𝑑𝑦
𝑥 𝛤𝑛 𝑦 +𝑏
 𝑐𝑜𝑠𝑏𝑥 1 𝑦 1 𝑦
𝑜𝑟, 𝑑𝑥 = 𝑦 𝑑𝑦 = 𝑑𝑦
𝑥 𝛤𝑛 𝑦 +𝑏 𝛤𝑛 𝑦 +𝑏

𝑐𝑜𝑠𝑏𝑥 1 𝑦
𝑜𝑟, 𝑑𝑥 = 𝑑𝑦
𝑥 𝑏 𝛤𝑛 1+( )

𝑐𝑜𝑠𝑏𝑥 𝑏 𝑢
𝑜𝑟, 𝑑𝑥 = 𝑑𝑢, 𝑝𝑢𝑡 𝑦/𝑏 = 𝑢
𝑥 2𝛤𝑛 1+𝑢

𝑐𝑜𝑠𝑏𝑥 𝑏 𝑝( )
𝑜𝑟, 𝑑𝑥 = 𝑑𝑝, 𝑝𝑢𝑡 𝑝 = 𝑢
𝑥 2𝛤𝑛 1+𝑝
( )
𝑐𝑜𝑠𝑏𝑥 𝑏 𝑝
𝑜𝑟, 𝑑𝑥 = 𝑑𝑝
𝑥 2𝛤𝑛 1+𝑝

𝑐𝑜𝑠𝑏𝑥 𝑏 𝜋 𝑦 𝜋
𝑜𝑟, 𝑑𝑥 = . ,[ 𝑑𝑦 = , 0 < 𝑛 < 1]
𝑥 2𝛤𝑛 𝑠𝑖𝑛𝑛 (𝑛 + 1)𝜋 (1 + 𝑦) 𝑠𝑖𝑛𝑛𝜋

𝑐𝑜𝑠𝑏𝑥 𝑏 𝜋
∴ 𝑑𝑥 = . , 0 < 𝑛 < 1.
𝑥 𝛤𝑛 2𝑐𝑜𝑠𝑛𝜋/2

Ex.8.Show that ∫ 𝑑𝑥 = . , 0 < 𝑛 < 1.

/

/ ( ( ))
Ex9: Show that ∫ = .
( / √
 Beta function:

Beta Function: The definite integral ∫ 𝑥 (1 − 𝑥) 𝑑𝑥 , 𝑚, 𝑛 > 0 is known as

first Eulerian integral is called beta function of 𝑚 & 𝑛 and expressed as

(i) 𝛽(𝑚, 𝑛) = ∫ 𝑥 (1 − 𝑥) 𝑑𝑥 , 𝑚, 𝑛 > 0.

Another form of beta function:

𝑦 𝑥
𝛽(𝑚, 𝑛) = 𝑑𝑦 = 𝑑𝑥 , ( 𝑝𝑢𝑡 𝑥 = 1/(1 + 𝑦) 𝑖𝑛 (𝑖))
(1 + 𝑦) (1 + 𝑥)

𝛽(𝑚, 𝑛) = 2 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝑑𝜃 = 2 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝑑𝜃 ,

( 𝑝𝑢𝑡 𝑥 = 𝑠𝑖𝑛 𝜃 𝑖𝑛 (𝑖))

𝛽(𝑚, 𝑛) = ∫ 𝑥 (𝑎 − 𝑥) 𝑑𝑥 , ( 𝑝𝑢𝑡 𝑥 = 𝑦/𝑎 𝑖𝑛 (𝑖)).

Properties of beta function:

𝛽(𝑚, 𝑛) = 𝛽(𝑛, 𝑚).

Relation Gamma & Beta function: 𝛽(𝑚, 𝑛) =

( )

𝛤𝑛 = 𝑎 ∫ 𝑒 𝑥 𝑑𝑥 (1)

Multiplying (1) by 𝑒 𝑎 and integrating with respect to 𝑎 from 0 𝑡𝑜 ∞, we get

𝛤𝑛 𝑒 𝑎 𝑑𝑎 = [𝑒 𝑎 𝑎 𝑒 𝑥 𝑑𝑥]𝑑𝑎

( )
or, 𝛤𝑚𝛤𝑛 = ∫ ∫ 𝑎 𝑒 𝑥 𝑑𝑥𝑑𝑎

( )
𝑜𝑟, 𝛤𝑚𝛤𝑛 = ∫ ∫ 𝑎 𝑒 𝑥 𝑑𝑎𝑑𝑥 [Interchanging order of integration]

( )
𝑜𝑟, 𝛤𝑚𝛤𝑛 = 𝑥 { 𝑎 𝑒 𝑑𝑎}𝑑𝑥

( )
𝑜𝑟, 𝛤𝑚𝛤𝑛 = ∫ 𝑥 𝑑𝑥 [comparing 𝑎 with (1 + 𝑥)in (1)]
( )

𝑜𝑟, 𝛤𝑚𝛤𝑛 = 𝛤(𝑚 + 𝑛) ∫ 𝑑𝑥 (ii)

( )
we know 𝛽(𝑚, 𝑛) = ∫ 𝑥 (1 − 𝑥) 𝑑𝑥 , ( 𝑝𝑢𝑡 𝑥 = , 𝑑𝑥 = − ( )
𝑑𝑦)

when 𝑥 = 0, 𝑡ℎ𝑒𝑛 𝑦 = ∞ and 𝑥 = 1, 𝑡ℎ𝑒𝑛 𝑦 = 0

1 𝑦 1
𝛽(𝑚, 𝑛) = − ( ) ( ) 𝑑𝑦
1+𝑦 1+𝑦 (1 + 𝑦)

𝑜𝑟, 𝛽(𝑚, 𝑛) = ∫ 𝑑𝑦 = ∫ 𝑑𝑥 (iii)

( ) ( )

From (ii) and (iii) we get

𝛤𝑚𝛤𝑛 = 𝛤(𝑚 + 𝑛)𝛽(𝑚, 𝑛)

𝛤𝑚𝛤𝑛
∴ 𝛽(𝑚, 𝑛) =
𝛤(𝑚 + 𝑛)
( ) ( )
( ) ( )
Ex:1 Show that ∫ 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜃𝑑𝜃 = ( ) , 𝑚 > −1, 𝑛 > −1.
( )

/ /
L.H.S= ∫ 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜃𝑑𝜃 = ∫ (𝑠𝑖𝑛 𝜃) (𝑐𝑜𝑠 𝜃) 𝑑𝜃

/ /
= ∫ (𝑠𝑖𝑛 𝜃) (1 − 𝑠𝑖𝑛 𝜃) 𝑑𝜃 𝑝𝑢𝑡 𝑧 = 𝑠𝑖𝑛 𝜃, 𝑑𝑧 = 2𝑎𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜃𝑑𝜃

1
= 𝑧 (1 − 𝑧) 𝑑𝑧
2

1
= 𝑧 (1 − 𝑧) 𝑑𝑧
2

1
= 𝑧 (1 − 𝑧) 𝑑𝑧
2
1 𝑚+1 𝑛+1
= 𝛽( , )
2 2 2
(𝑚 + 1) (𝑛 + 1)
𝛤( )𝛤( ) 𝛤𝑚𝛤𝑛
= 2 2 , [∴ 𝛽(𝑚, 𝑛) = ]
(𝑚 + 𝑛 + 2) 𝛤(𝑚 + 𝑛)
2𝛤( )
2
(𝑚 + 1) (𝑛 + 1)
𝛤( )𝛤( )
∴ 𝑠𝑖𝑛 𝜃 𝑐𝑜𝑠 𝜃𝑑𝜃 = 2 2
(𝑚 + 𝑛 + 2)
2𝛤( )
2
 ( )
( )√
∫ 𝑠𝑖𝑛 𝜃𝑑𝜃 = ∫ 𝑐𝑜𝑠 𝜃𝑑𝜃 = ( ) .
( )

Ex:2; Show that 𝛤𝑛 𝛤(1 − 𝑛) = , 0 < 𝑛 < 1.

we know 𝛽(𝑚, 𝑛) = ∫ 𝑥 (1 − 𝑥) 𝑑𝑥 , ( 𝑝𝑢𝑡 𝑥 = , 𝑑𝑥 = − ( )

𝑑𝑦)

when 𝑥 = 0, 𝑡ℎ𝑒𝑛 𝑦 = ∞ and 𝑥 = 1, 𝑡ℎ𝑒𝑛 𝑦 = 0

1 𝑦 1
𝛽(𝑚, 𝑛) = − ( ) ( ) 𝑑𝑦
1+𝑦 1+𝑦 (1 + 𝑦)

𝑜𝑟, 𝛽(𝑚, 𝑛) = ∫ 𝑑𝑦
( )

𝑜𝑟, ( )
=∫ 𝑑𝑦
( )

put 𝑚 + 𝑛 = 1 𝑜𝑟 𝑚 = 1 − 𝑛
( )
( )
=∫ 𝑑𝑦
( )

we know that the standard formula of improper integral

𝑦
𝑑𝑦 = 𝜋𝑐𝑜𝑠𝑒𝑐𝑛𝜋, 0 < 𝑛 < 1
(1 + 𝑦)

∴ 𝛤𝑛𝛤(1 − 𝑛) = 𝜋𝑐𝑜𝑠𝑒𝑐𝑛𝜋 = ,0 < 𝑛 < 1

Error function:

Error function is defined as erf(𝑥) = ∫ 𝑒 𝑑𝑟. This integral also known as

√
probability integral in statistics. The probability integral is encountered in many
branches of applied mathematics e.g., probability theory, the theory of error, the
theory of heat conduction and varies branches of mathematical physics.

2
erf(0) = 𝑒 𝑑𝑟 = 0
√𝜋
2
erf(−𝑥) = 𝑒 𝑑𝑟 = − erf(𝑥)
√𝜋
∴ erf(−𝑥) is an odd function
 √
Also erf(∞) = ∫ 𝑒 𝑑𝑟 = . =1
√ √

This show the maximum probability is 1 and the complementary Error function of
𝑥 is defined as erf c(𝑥) = 1 − erf (𝑥) = ∫ 𝑒 𝑑𝑟
√

∴ erf c(0) = 1, erf c(∞) = 0.