1
$\begingroup$

The lower incomplete gamma function for positive $s$ is defined by the integral $$ \gamma(s,x)=\int_0^{x} t^{s-1} e^{-t} dt. $$ Taylor expansion of the exponential function and term by term integration give the following expansion $$ \gamma(s,x)=\sum_{n=0}^\infty \frac{(-1)^n x^{n+s}}{n! (n+s)} $$ Here $\gamma(s,x)$ can be analytically continued for complex $s$ except some singular points. Does the above expansion still hold in this case? Especially I'd like to know if the relation holds for $s < 0$. Each term of the series is well defined for $s < 0$. Is this enough to ensure the validity of the relation or does it need more arguments?

$\endgroup$
  • $\begingroup$ You need more arguments. If you could show that the sum is an analytic function of $s$ for $s \neq 0, -1, -2, \ldots$ then the equality would hold for $s \neq 0, -1, -2, \ldots$ by analytic continuation. For such an argument one could probably make use of the dominated convergence theorem. $\endgroup$ – Antonio Vargas Mar 13 '14 at 16:59
1
$\begingroup$

From http://dlmf.nist.gov/8.7.E3 we have the series expansion $$\Gamma(s,x) = \Gamma(s) - \sum_{n=0}^\infty \frac{(-1)^n x^{n+s}}{n! (n+s)}, \qquad s \ne 0, -1, -2, \ldots $$ Combine this with the relation for the gamma functions (http://dlmf.nist.gov/8.2.E3) $$\gamma(s,x) + \Gamma(s,x) = \Gamma(s).$$ Therefore the series expansions remains valid for all non-integer $s<0$ $$ \gamma(s,x) = \sum_{n=0}^\infty \frac{(-1)^n x^{n+s}}{n! (n+s)}, \qquad s \ne 0, -1, -2, \ldots $$ Another route is via Tricomi's entire incomplete gamma function $\gamma^{*}$, see http://dlmf.nist.gov/8.7.E1.

$\endgroup$
  • $\begingroup$ Thank you for the answer. Ok, the expansion is right for complex $s$. I understand the $\gamma^*$ approach. By the way, do you have any comment about my last question? $\endgroup$ – asofas Mar 13 '14 at 14:43
  • $\begingroup$ I guess with validity you mean (well-)defined. This is necessary but you still need the convergence of the sum. $\endgroup$ – gammatester Mar 13 '14 at 15:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.