Polygamma function explained

C defined as the th derivative of the logarithm of the gamma function:\psi ^{(m)} (z) : = (d ^{m})/(d z ^{m}) \psi(z) = (d ^{m+1})/(d z ^{m+1}) ln \Gamma(z).

Thus

\psi ^{(0)} (z) = \psi(z) = (\Gamma'(z))/(\Gamma(z))

holds where is the digamma function and is the gamma function. They are holomorphic on

C \backslash Z _{\le0} . At all the nonpositive integers these polygamma functions have a pole of order . The function is sometimes called the trigamma function.

Integral representation

When and, the polygamma function equals

\begin{align} \psi ^{(m)} (z) & = ( -1 ) ^{m+1} \int ^{inf ty} _{0} (t ^{m} e ^{-zt})/(1- e ^{-t}) d t \\ & = - \int ^{1} _{0} (t ^{z-1})/(1- t) ( ln t) ^{mdt\\ &=} ( -1 ) ^{m+1} m!\zeta(m +1 ,z) \end{align}

where

\zeta(s,q) is the Hurwitz zeta function.

This expresses the polygamma function as the Laplace transform of . It follows from Bernstein's theorem on monotone functions that, for and real and non-negative, is a completely monotone function.

Setting in the above formula does not give an integral representation of the digamma function. The digamma function has an integral representation, due to Gauss, which is similar to the case above but which has an extra term .

Recurrence relation

It satisfies the recurrence relation

\psi ^{(m)} (z +1 ) = \psi ^{(m)} (z) + (( -1 ) ^{mm!})/(z ^{m+1})

which – considered for positive integer argument – leads to a presentation of the sum of reciprocals of the powers of the natural numbers:

(\psi ^{(m)} (n))/(( -1 ) ^{m+1} m!) = \zeta( 1+ m) - \sum ^{n -1} _{k =1} (1)/(k ^{m+1}) = \sum ^{inf ty} _{k = n} (1)/(k ^{m+1}) m \ge 1

and

\psi ^{(0)} (n) = - \gamma + \sum ^{n -1} _{k =1} (1)/(k)

for all

n \in N , where \gamma is the Euler–Mascheroni constant. Like the log-gamma function, the polygamma functions can be generalized from the domain uniquely to positive real numbers only due to their recurrence relation and one given function-value, say, except in the case where the additional condition of strict monotonicity on R ^{+} is still needed. This is a trivial consequence of the Bohr–Mollerup theorem for the gamma function where strictly logarithmic convexity on R ^{+} is demanded additionally. The case must be treated differently because is not normalizable at infinity (the sum of the reciprocals doesn't converge).

Reflection relation

( -1 ) ^{m} \psi ^{(m)} ( 1- z) - \psi ^{(m)} (z) = \pi (d ^{m})/(d z ^{m}) \cot{\pi z} = \pi ^{m+1} (P _{m(\cos{\pi} z)/()}{\sin) ^{m+1} (\pi z)}

where is alternately an odd or even polynomial of degree with integer coefficients and leading coefficient . They obey the recursion equation

\begin{align} P _{0(x)} & = x \ P _{m+1} (x) & = - \left( (m +1 )xP ^{2 \right)P'} _{m(x)\right).\end{align}}

Multiplication theorem

The multiplication theorem gives

k ^{m+1} \psi ^{(m)} (kz) = \sum ^{k -1} _{n =0} \psi ^{(m)} \left(z + (n)/(k) \right) m \ge 1

and

k \psi ^{(0)} (kz) = k ln {k} + \sum ^{k -1} _{n =0} \psi ^{(0)} \left(z + (n)/(k) \right)

for the digamma function.

Series representation

The polygamma function has the series representation

\psi ^{(m)} (z) = ( -1 ) ^{m+1} m! \sum ^{inf ty} _{k =0} (1)/((z + k) ^{m+1})

which holds for integer values of and any complex not equal to a negative integer. This representation can be written more compactly in terms of the Hurwitz zeta function as

\psi ^{(m)} (z) = ( -1 ) ^{m+1} m! \zeta (m +1 ,z).

This relation can for example be used to compute the special values^[1]

\psi ^{(2n-1)} \left( (14 \right))/(=) (4 ^{2n-1})/(2 n) \left(\pi ^{2n} ( 2 ^{2n} -1 )|B _{2n} | +2 ( 2 n)!\beta( 2 n)\right); \psi ^{(2n-1)} \left( (34 \right))/(=) (4 ^{2n-1})/(2 n) \left(\pi ^{2n} ( 2 ^{2n} -1 )|B _{2n} | -2 ( 2 n)!\beta( 2 n)\right); \psi ^{(2n)} \left( (14 \right))/(=) -2 ^{2n-1} \left(\pi ^{2n+1} |E _{2n} | +2 ( 2 n)!( 2 ^{2n+1} -1 )\zeta( 2 n +1 )\right); \psi ^{(2n)} \left( (34 \right))/(=) 2 ^{2n-1} \left(\pi ^{2n+1} |E _{2n} | -2 ( 2 n)!( 2 ^{2n+1} -1 )\zeta( 2 n +1 )\right).

Alternately, the Hurwitz zeta can be understood to generalize the polygamma to arbitrary, non-integer order.

One more series may be permitted for the polygamma functions. As given by Schlömilch,

(1)/(\Gamma(z)) = z e ^{\gamma} \prod ^{inf ty} _{n =1} \left( 1 + (z)/(n) \right) e ^{- (z)/(n)} .

This is a result of the Weierstrass factorization theorem. Thus, the gamma function may now be defined as:

\Gamma(z) = (e ^{-\gamma})/(z) \prod ^{inf ty} _{n =1} \left( 1 + (z)/(n) \right) ^{-1} e ^{(z)/(n)} .

Now, the natural logarithm of the gamma function is easily representable:

ln \Gamma(z) = - \gamma z - ln (z) + \sum ^{inf ty} _{k =1} \left( (z)/(k) - ln \left( 1 + (z)/(k) \right) \right).

Finally, we arrive at a summation representation for the polygamma function:

\psi ^{(n)} (z) = (d ^{n+1})/(d z ^{n+1}) ln \Gamma(z) = - \gamma \delta _{n0} - (( -1 ) ^{n} n!)/(z ^{n+1}) + \sum ^{inf ty} _{k =1} \left( (1)/(k) \delta _{n0} - (( -1 ) ^{n} n!)/((k + z) ^{n+1}) \right)

Where is the Kronecker delta.

\Phi( -1 , m +1 , z) = \sum ^{inf ty} _{k =0} (( -1 ) ^{k})/((z + k) ^{m+1})

can be denoted in terms of polygamma function

\Phi( -1 , m +1 , z) = (1 {( -2 ))/(^{m+1}) m!}\left(\psi ^{(m)} \left( (z)/(2) \right) - \psi ^{(m)} \left( (z +1)/(2) \right)\right)

Taylor series

The Taylor series at is

\psi ^{(m)} (z +1 ) = \sum ^{inf ty} _{k =0} ( -1 ) ^{m+k+1} ((m + k)!)/(k!) \zeta (m + k +1 ) z ^{k} m \ge 1

and

\psi ^{(0)} (z +1 ) = - \gamma + \sum ^{inf ty} _{k =1} ( -1 ) ^{k+1} \zeta (k +1 ) z ^{k}

which converges for . Here, is the Riemann zeta function. This series is easily derived from the corresponding Taylor series for the Hurwitz zeta function. This series may be used to derive a number of rational zeta series.

Asymptotic expansion

These non-converging series can be used to get quickly an approximation value with a certain numeric at-least-precision for large arguments:^[2]

\psi ^{(m)} (z) \sim ( -1 ) ^{m+1} \sum ^{inf ty} _{k =0} ((k + m -1 )!)/(k!) (B _{k})/(z ^{k+m}) m \ge 1 and\psi ^{(0)} (z) \sim ln (z) - \sum ^{inf ty} _{k =1} (B _{k})/(k z ^{k})

where we have chosen, i.e. the Bernoulli numbers of the second kind.

Inequalities

The hyperbolic cotangent satisfies the inequality

(t)/(\operatorname{coth} 2) (t)/(2) \ge 1 , and this implies that the function(t ^{m})/(1- e ^{-t}) - \left(t ^{m-1} + (t ^{m})/(2) \right) is non-negative for all and . It follows that the Laplace transform of this function is completely monotone. By the integral representation above, we conclude that( -1 ) ^{m+1} \psi ^{(m)} (x) - \left( ((m -1 )!)/(x ^{m}) + (m!)/(2 x ^{m+1}) \right) is completely monotone. The convexity inequality implies that\left(t ^{m-1} + t ^{m\right)} - (t ^{m})/(1- e ^{-t}) is non-negative for all and, so a similar Laplace transformation argument yields the complete monotonicity of\left( ((m -1 )!)/(x ^{m}) + (m!)/(x ^{m+1}) \right) - ( -1 ) ^{m+1} \psi ^{(m)} (x). Therefore, for all and,((m -1 )!)/(x ^{m}) + (m!)/(2 x ^{m+1}) \le ( -1 ) ^{m+1} \psi ^{(m)} (x) \le ((m -1 )!)/(x ^{m}) + (m!)/(x ^{m+1}) . Since both bounds are strictly positive for x> 0 , we have:ln \Gamma(x) is strictly convex.

• For

m =0 , the digamma function, \psi(x) = \psi ^{(0)} (x) , is strictly monotonic increasing and strictly concave.

• For

m odd, the polygamma functions, \psi ^{(1)} ,\psi ^{(3)} ,\psi ^{(5)} ,\ldots , are strictly positive, strictly monotonic decreasing and strictly convex.

• For

m even the polygamma functions, \psi ^{(2)} ,\psi ^{(4)} ,\psi ^{(6)} ,\ldots , are strictly negative, strictly monotonic increasing and strictly concave. This can be seen in the first plot above.

Trigamma bounds and asymptote

For the case of the trigamma function (

m =1 ) the final inequality formula above for x> 0 , can be rewritten as:(x + (12)/())/(x ^{2}) \le \psi ^{(1)} (x)\le (x +1)/(x ^{2}) so that for x\gg 1 : \psi ^{(1)} (x) ≈ (1 x)/() .

See also

• Factorial
• Gamma function
• Digamma function
• Trigamma function
• Generalized polygamma function

References

• Book: Milton. Abramowitz. Irene A.. Stegun. Handbook of Mathematical Functions. 1964. Dover Publications. New York. 978-0-486-61272-0. https://personal.math.ubc.ca/~cbm/aands/page_260.htm. Section 6.4.

Notes and References

1. K. S. . Kölbig. 1996. J. Comput. Appl. Math.. 75. 1. 43–46. The polygamma function psi^k(x) for x=1/4 and x=3/4. 10.1016/S0377-0427(96)00055-6. free.
2. J.. Blümlein. Comp. Phys. Comm.. 2009. 180. 2218–2249. 10.1016/j.cpc.2009.07.004. Structural relations of harmonic sums and Mellin transforms up to weight w=5. 11 . 0901.3106. 2009CoPhC.180.2218B .