Professional Documents
Culture Documents
Contents
Page
Mathematical Properties. . . . . . . . . . . . . . . . . . . . 255
6.1. Gamma Function. . . . . . . . . . . . . . . . .
. . . 255
6.2. Beta Function . . . . . . . . . . . . . . . . . .
. . . 258
6.3. Psi (Digamma) Function. . . . . . . . . . . . . .
. . . 258
6.4. Polygamma Functions. . . . . . . . . . . . . . .
. . . 260
6.5. Incomplete Gamma Function. . . . . . . . . . . . . . . 260
6.6. Incomplete Beta Function. . . . . . . . . . . . . . . . 263
Numerical Methods . . . . . . . . . . . . . . . . . . . . . . 263
6.7. Use and Extension of the Tables. . . . . . . . . . . . . 263
6.8. Summation of Rational Series by Means of Polygamma Func-
tions. . . . . . . . . . . . . . . . . . . . . . . . . 264
References. . . . . . , . . . . . . . , . . . . . . . . . . . . 265
Table 6.1. Gamma, Digamma and Trigamma Functions (1 5 s l 2 ) . . 267
r(x),~n r(x), +(z),+'(z), ~=i(.oo5)2, IOD
Table 6.2. Tetragamma and Pentagamma Functions (1 5 x 5 2 ) . . . 27 1
+"(x), $J3'(2),~=1(.01)2, 1OD
Table 6.3. Gamma and Digamma Functions for Integer and Half-
Integer Values ( l l n 5 1 0 1 ) . . . . . . . . . . . . . . . . . . 272
r(n), 11s +(n), IOD
l/r(n), 9s n!/[(2?r)h"+3]~",8D
r(n+$), 8s Inn-+@), 8D
n=1(1.)101
Table 6.4. Logarithms of the Gamma Function (1 I n 5 101). . . . . . 274
loglo r(n>, 8s log10 r(n+$), 8s
log10 r(n+#), 8 s In r(n)-(n-$) lnn+n, 8D
log10 r(n+$), 8s
n= l(1) 101
Page
Table 6.5. Auxiliary Functions for Gamma and Digamma Func-
. . 276
6.1.4 -=-s1
r(z) 2,
Hankel's Contour Integral
i
c
(-t)-'e-'dt
(k<l =)
6.1.10 r (n+ 3) =1.5.9.13.4". . (4n-3) r(t>
r(+)=3.6256099082. . .
The path of integration C starts at + QD on the
1.4-7.10. . . (3n-2)
real axis, circles the origin in the counterclockwise 6.1.11 r(n+#)=
3" r (4)
direction and returns to the starting point.
Factorial and II Notations r($)=2.67893 85347 . . .
6.1.5 n(z)=z!=r(z+i) 1-3-5-7... (2n-1)
6.1.12 r(n+$) =
2" r (3)
Integer Values
6.1.6 r(n+1)=1.2.3 . . . (n-l)n=n! 2.5.8-11.. . (3n-1)
6.1.13 l"(n+#)=
3" r(3)
6.1.7
lim -=o=
1 1 (n=O, 1, 2, . . .) r(#)=i.3aii 79394. . .
z+, r(-z) (-n-l)!
6.1.8
Fractional Values 6.1.14 r(n+i)= 3.7.11.15.4". . (4n-1) Ut)
I-()) = 2 s me-12dt=&=1.77245 38509 . . . =(-3)!
0 r($)=i.22541 67024 . . .
*See page 11. 255
256 GAMMA FUNCTION AND RELATED FUNCTIONS
Recurrence Formulas 6.1.29 r(i~)r(-iy)=ir(i~)iz=9r
6.1.15 r(z+i)=Zr(Z)=Z!=Z(Z-i)! y sinh 7ry
6.1.16
r(n+~)=(n-il+z)(n-2+~) . . . (i+z)r(i+z)
= (n- 1+ z)! 6.1.31 r (1 +iy)r (1-iy)= Ir (1 +iY) (L“ ?/
sinh ry
=(n-l+z)(n-2+2). . . (l+z)z!
Reflection Formula
6.1.17 r(z>r(i-z)=-zr(-z)r(z)=t a c 7rz
-
=J=, tz-1
-dt (O< 9 2 <1)
6.1.33
Power Series
k ck
1 1.00000 00000 000000
2 0.57721 56649 015329
Binomial Coefficient
3 -0.65587 80715 202538
4 -0.04200 26350 340952
5 0. 16653 86113 822915
6 -0.04219 77345 555443
Pochhammer’sSymbol 7 -0.00962 19715 278770
6.1.22 8 0.00721 89432 466630
@lo= 1, 9 -0.00116 51675 918591
(2),=2(2+1)(2+2) . . . (z+n-l)=- r(z+n) 10 -0.00021 52416 741149
r (2) 11 0.00012 80502 823882
12 -0.00002 01348 547807
Gamma Function in the Complex Plane
13 -0.00000 12504 934821
6.1.23 r@)=r);
In r(Z)=In r(z) 14 0.00000 11330 272320
15 -0.00000 02056 338417
16 0.00000 00061 160950
6.1.24 arg r(z+l)=arg r(z)+arctan 17 0.00000 00050 020075
X
18 -0.00000 00011 812746
19 0.00000 00001 043427
20 0.00000 00000 077823
21 -0.00000 00000 036968
22 0.00000 00000 005100
23 -0.00000 00000 000206
24 -0.00000 00000 000054
25 0.00000 00000 000014
26 0.00000 00000 000001
2 The coefficients ck are from H. T. Davis, Tables of
higher mathematical functions, 2 vols., Principia Press,
Bloomington, Ind., 1933, 1935 (with permission) ; with
corrections due to H. E. Salzer.
GAMMA FUNCTION AND RELATED F"CX'I0NS 257
Polynomial Approximations' Error Term for Asymptotic Expansion
where
K(z)=upperU l bound(z2/(u2+z3)
O I
bl=-. 57719 1652 bs=-. 75670 4078
bs= .98820 5891 be= .48219 9394
b3= -. 89705 6937 b7= -. 19352 7818 For z real and positive, R, is less in absolute value
bq= ,91820 6857 bs= .03586 8343 than the first term neglected and has the same
ign.
Stirling's Formula
6.1.43
6.1.37
9th r(iy)=%?ln r(-iy)
+z m
B2m
2rn(2m-1)2*-'
(z+m in larg z~<T) 6.1.46
From C. Hmtings, Jr., Approximations for digital as z+m along any curve joining z=O and Z= m,
computers, Princeton Univ. Press, Princeton, N.J., 1955 providingz# --a, ---a-1, . . . ; zf --b, -b-1,
(with permission). . . . .
- (2n)!
--- 1 2n r(n++)
22n(n!)2-~ (n)=rtr(n+i)
- 1 1
-&j [I-%+=*-
1
* * *
1 6.2. Psi function.
FIGURE
y = $(z) = d In r kc)/&
(n+ OJ 1
Some Definite Integrals
6.1.50 Integer Values
9
Psi Function in the Complex Plane (Y+OJ)
-
6.3.9 +GI=*(z> Extremaoof r(z) - Zeros of $(z)
6.3.10
9+(iy>=W+(-iy)=W+(l +iy)=W+(l -iy)
6-3-11 coth xy
Y+(iy)=&/-'+#,~ I
6.3.14 +(l+~)=-r+C(-l)"~(n)~~-'
n-2
(Iz1<1)
Zo=1.46163 21449 68362
6.3.15 r(xo)=.88560 31944 10889
+( 1 +2) =&-1- &€ cot e-(1-9) -'+ 1-7
6.3.20 zn=-n+(ln n)-'+o[(ln n)-*]
-n-1
5It(2n+ 1) -11ZSA (Iz I <2) Definite Integrals
6.3.16 6.3.21
(~#-1,-2,-3, ...)
6.3.17
9+(l+iy)=l-r-- 1
l+y2
+g(- l)"+'[r(2n+1) -l]y2'
n=1
(IYl<2)
= -r+ y2 c n-'(n*+yS)
OD
a-1
-1
(- Y<
OJ -1
Asymptotic Formulae
6.3.18
1
-In z-s-n-l c-
= Bz,
2nz2"
=In z----1 1 1 1
22 1 2 9 + 1 2 0 2 4 - ~ 6+ . . . 6From W. Sibagaki, Theory and applications of the
gamma function, Iwanami Syoten, Tokyo, Japan, 1952
(z+- in lergzl<*) (with permission).
260 GAMMA FUNCTION AND RELATED FUNCTION8
* 6.4.10
$(") (2)= (-1) "+In!
-(2 +k) -*-I
$(n)(z)J(n=OJIJ . . .), is a single valued analytic k-0
function over the rnt.irc complex plane save at (z#0,-1,-2, . . .)
the points z=-m(m=0,1,2, . . . ) where it pos-
sesses poles of order (n+ 1). Asymptotic Formulas
6.4.11
Integer Values
6.4.2
$(")(l)=(- l)"+'n!{(n+1) (n=€,2,3, . .)
6.4.3 6.4.12
1 1- +---+---
1 - 1 1 1
$' (2) --+ +...
z 22' 6z3 30z6 422' 302'
(z+= in I arg zl<r)
6.4.13
Fractional Values
6.4.4
$'"'(*)=(-1)"+'n!(2"+'-l){(n+l)
(n=1,2, . . . )
6.4.14
6.4.5 $'(n++)=+$-4 2(2k- 1)-*
1;-1
2 3 2 1 +-'--+--
$'B)(z) "-+7+---
23 2
4 3 1
zS 2' 32
0
213
, ..
Recurrence Formula
(z+w in I arg z I<r)
=J-e-tta-ldt
6.5.3
r(a, r) = r(a>-r(a, r)
6=1, n=O z
*See page n.
262 GAMMA FUNCTION AND RELATED F"CT1ONS
6.5.5 6.5.16
Probability Integral of the +Distribution
6.5.17
6.5.18
6.5.6
(Pearson's Form of the Incomplete Gamma Function)
6.5.19
6.5.20
Recurrence Formulas
m
6.5.7 C(z,a)=l tu-1 cos t dt (L@'a<l)
9e-"
6.5.21 P(a+l, z)=P(a, z)----
m
r(a+l>
6.5.8 S(z,a)=$, ta-l sin t dt (9'a<l)
6.5.22 y (a+1,z)=uy(a,z)
6.5.9 e-'
6.5.23 V*(u-l,z) =m*(u,z)+-r (a)
nm
6.5.24
6.5.11
=1-e,,- (2)e-2
Series Developmente
For relation to the Poisson distribution, see
26.4. 6.5.29
6.5.14 r*(-n, z)=z"
Continued Fraction
6.5.31
niitl sign I?,(a,z) =sign u,<+,(a,z). For binomial distribution, see 26.1.
Recurrencc Formulas
(z+m in I nrg +
<)zlr 6.6.8 B,(a,b)=a-'~'c"F(a,l-b; a + l ; Z)
Numeric:a1 Methods
6.7. Use and Extension of the Tables
The crror of liiicar intrrpolation in the table of
Example 1. Compute r(6.38) to 8s. Using tlic function f 2 is smaller than lo-' in this region.
the r ~ w i r r ~ ~ i irchtioii
cc 6.1.16 niitl Table 6.1 wc Hence, f2(56.38)= .92041 67 and In I'(56.38) =
1 1 avc, 169.85497 42.
r (6.38)= [(5.38)(4.38)(3.38)(2.33)(1.38)]r (1.38) Direct interpolation in Table 6.4 of log,, r(n)
= 232.43671. climiiiatcs tlic necessity of employing logarithms.
HOWCVP~, tlic rrror of liiicar intcrpolation is .002 so
Example 2. Compute In r(56.38), iisiiig Table tltnt log,, r(n) is obtained with a rclativc error
6.4 niid liiicnr iiitrrpolation iii j...
\\-e liavc of 10-5.
In r(56.38) = (56.38-3) In (56.38)- (56.38) *See page 11.
+j 2 (56.38)
264 GAMMA FUNCTION AN RELATED FUNCTIONS
Example 3. Compute $(6.38) to 8s. Using the 6.8. Summation of Rational Series by Means
recurrence relation 6.3.6 and Table 6.1. of Polygamma Functions
we have
a1=1, a2=3, as=*, al=*, &=-l, *=#. i e=-,
-i6 -i i
Hence, al=-,6 &=-12 ' a 4 = 3
Thus,
a1=i, az=-i, aa=2i, a,=-2i,
8= -)$(2) +$(13) -#$(It) =.047198.
and therefore
Example 10.
m 1
s=- --z [$(1 +i)-$(1 -ill +ai [$(1+2i) -$(1-2i)l.
6
References
Texts Tables
[6.1] E. Artin, Einfiihrung in die Theorie der Gamms- [6.9] A. Abramov, Tables of r(z) for complex argu-
I n
[6.14] National Bureau of Standards, Tables of nl and [6.18] E. S. Pearson, Table of the logarithms of the com-
r(n+& for the first thousand values of n, Ap- plete r-function, arguments 2 to 1200, Tracts for
plied Math. Series 16 (U.S.Government Printing Computers No. VI11 (Cambridge Univ. Press,
O5ce, Washington, D.C., 1951). nf, 16S;r(n+&, Cambridge, England, 1922). Loglo r(p), p=2(.1)
8s. 5(.2)70(1)1200, 10D.
[6.15] National Bureau of Standards, Table of Coulomb
wave functions, vol. I, pp. 114-135, Applied [6.19] J. Peters, Ten-place logarithm tables, vol. I, Ap-
Math. Series 17 (U.S. Government Printing
pendix, pp. 58-68 (Frederick Ungar Publ. Co.,
O5ce, Washington, D.C., 1952).
New York, N.Y., 1957). nl, n=1(1)60, exact;
9[ryi+is)/r(1 +is],9 =o(.oo5)2 (.oi)6 (.02)1o(.1 (n!)-’, n=1(1)43, 54D; Log,o(nl), n=1(1)1200,
20 (.2)60(.5)1 10,lOD; apg r (1+is),s = O(.el)1 (.02) 18D.
3 (.05)10(.2)20(.4)30(.5)85, 8D.
(6.20) J. P. Stanley and M. V. Wilkes, Table of the recip-
[6.16] National Bureau of Standards, Table of the gamma rocal of the gamma function for complex argu-
function for complex arguments, Applied Math.
ment (Univ. of Toronto Press, Toronto, Canada,
Series 34 (U.S. Government Printing O5ce,
Washington, D.C., 1954). 1950). Z= -.5( .01).5, y=O(.Ol)l, 6D.