Main Index
Mathematical Analysis
Special Functions
Subject Index
comment on the page
Gamma function can be defined in many various ways. The following definition goes back to Euler (1729) and from that reason the defined function is also called Euler Gamma function or Euler function of the second kind .
(1) |
The starting point of Euler’s definition was the identity which he used to extend the factorial function from the natural numbers to all real numbers . Gauß introduced the notation for . Thus for positive integers . The name and notation go back to A.-M. Legendre (1809) (Cajori vol. 2, page 271). Legendre notation prevailed in France and later in the rest of world, despite the fact that Gauß “Pi” notation is more natural.
The Gamma function can be continued to the whole complex plane and it is there regular. As a function of complex variable has no zeros and a pole of order one at each of the negative integers and at zero. The residue at , is .
This is graph of the Gamma function for real values of :
The minimum of in interval is and it is attained at . The local minima of converge to 0 for .
This is the graph of the absolute value of the Gamma function
The behavior of for large is described in the so-called Stirling’s formula: If is fixed then the equality
(2) |
holds uniformly in the range
(3) |
from which the point and the poles of are cut off through small circles. The branches of and should be taken in such a way that their values are real at real positive . This gives the following estimate:
If and are such that then for we have
(4) |
for . The O-constant depends on and .
The first form of Stirling formula was found by de Moivre in the form where is a constant, the value of the constant was found by Stirling showing that . 1
A simple form of Stirling formula can be proved easily
Combined with the trapezoidal rule we get
for some constant . More precise form of Stirling’s formula is
(5) |
With the error estimates we have
(6) |
or
(7) |
etc.
Euler found the following convergent product approximation for non-integral values
(8) |
This is the graph of the real part of the Gamma function
This is the graph of the imaginary part of the Gamma function
Euler and Gauß defined the Gamma function originally also using the equivalent approach .
(9) |
Weierstrauß found the expression .
(10) |
where is the Euler-Mascheroni constant.This is an entire function in the complex plane.
The graph of for real values of the argument is:
Integration per parts yields from (1) the important functional equation
(11) |
(12) |
Consequently, if then
(13) |
For , the important connection to the factorial function easily follows
(14) |
Weierstrauß relation (10) implies
(15) |
This gives the relations (the first of them is called the Euler's reflection formula)
(16) |
(17) |
(18) |
(19) |
Relation (16) implies for that
(20) |
and for and we get
Note that no simple expression is known for , and . However, it was proved that these numbers are transcendental ( the first one by Le Lionnais [1] , p.46 in 1983 and two others by Chudnovsky [2] ,p. 308] in 1984 respectively).
Borwein and Zucker [3] showed how to reduce the evaluation of the gamma function at rational values by expressing gamma functions at rational values in terms of elliptic integrals. For instance
We also have
Legendre discovered the following duplication formula
(21) |
which was by Gauß extended to Gauß-Legendre multiplication formula
(22) |
In the above mentioned Gauß Pi notation the multiplication formula has the form
Since for , we can take the logarithm to get
We also have
(23) |
for , . The O-constant depends on .
A natural question in what extend the functional equation (12) determines the gamma function. The answer was given by the Bohr and Mollerup [4] :
There is a unique function such that is convex and , and .
The formula for the surface area of the -dimensional sphere is
and its volume is
1 | In the proof that the Wallis formula is instrumental. |
[1] | Le Lionnais, F. (1979). Les nombres remarquables. Paris: Hermann. |
[2] | Chudnovsky, G. V. (1984). Contributions to the theory of transcendental numbers. Providence, RI: Amer. Math. Soc. |
[3] | Borwein, J. M., & Zucker, I. J. (1992). Fast evaluation of the gamma function for small rational fractions using complete elliptic integrals of the first kind. IMA J. Numer. Anal., 12(4), 519-526. |
[4] | Bohr, H., & Mollerup, J. (1922). Laerebog i matematisk analyse, vol. 3 Graenseprocesser. (Danish). Kopenhagen: J. Gjellerup. |
Cite this web-page as:
Štefan Porubský: Gamma Function.