Loading Web-Font TeX/Math/Italic
Logo

2.2 Connecting "h" to the Gamma Function

When the integration interval is set from 0 to \infty , the gamma integral g\left( s+1, -1,u \right) becomes \Gamma \left( s+1 \right). If s is a nonnegative integer, it defines the factorial function s!. If s is not an integer, due to its recursive regularity, all gamma functions can be calculated as a function of \Gamma \left( \bmod (s,1) \right), where 0<\bmod \left( s,1 \right)<1.

\begin{equation*} \Gamma \left( s+1 \right)= \begin{cases} \Gamma \left( \bmod \left( s,1 \right) \right)\prod\limits_{i=0}^{\left\lfloor s \right\rfloor }{\left( s-i \right)}, &\text{if $s>-1$ & $s \notin {{\mathbb{Z}}^{+}_{0}}$;} \\ \Gamma \left( \bmod \left( s,1 \right) \right)\prod\limits_{i=0}^{\left| \left\lceil s \right\rceil \right|-1}{\left( s+1+i \right)^{-1}}, &\text{if $s<-1$ & $s \notin {{\mathbb{Z}}^{-}}$;} \\ s!, &\text{if $s\in {{\mathbb{Z}}^{+}_{0}}$.} \end{cases} \tag{2.3} \end{equation*}


[Examples of (2.3)]

For any non-integer s, the remainder for one, \bmod (s,1), is bounded within the interval \left( 0,1 \right). Thus,all gamma functions can be calculated if \Gamma \left( \bmod (s,1) \right) is known. Given that the "h" function is derived from factorizing the gamma integral, the investigation of "h" will focus on the same integral limit as \Gamma \left( \bmod (s,1) \right) .

An identity of the "h" function can be respecified from (2.2)

\begin{equation*} h_{k-r}^{c}={\exp(-c)}\sum\limits_{i=0}^{\infty }{\frac{{{c}^{i}}}{i!\left( k+1+i-r \right)}}. \tag{2.4} \end{equation*}


[Proof for (2.4)]


Here, k is a non-negative integer and 0<r<1. Performing long division for {1}/{\left[ \left( k+1+i \right)-r \right]}, we can derive

\begin{equation*} h_{k-r}^{c}={\exp(-c)}\sum\limits_{i=0}^{\infty }{\left\{ \frac{{{c}^{i}}}{i!}\left[ \sum\limits_{j=0}^{\infty } {\frac{{{r}^{j}}}{{{\left( k+i+1 \right)}^{j+1}}}} \right] \right\}}. \tag{2.5} \end{equation*}


[Proof for (2.5)]

If we rearrange (2.5) and sum all the terms by i given j, we derive

\begin{equation*} h_{k-r}^{c}=\frac{\exp(-c)}{{{c}^{k+1}}}\sum\limits_{i=0}^{\infty }{{{r}^{i}}{{I}^{\left( i \right)}}\left( c,k \right)}, \tag{2.6} \end{equation*}

where \begin{equation*} {{I}^{\left( 0 \right)}}\left( c,k \right)=\int{{{c}^{k}}\exp \left( c \right) dc}, \end{equation*}


and
\begin{equation*} {{I}^{\left( i \right)}}\left( c,k \right)=\int{\frac{^{(i)}{\iddots}\,\int{\frac{{{I}^{\left( 0 \right)}} \left( c,k \right)}{c}}dc}{c}}dc. \end{equation*}
[Proof for (2.6)]

We can further simplify {{I}^{\left( n \right)}}\left( c,k \right) as

\begin{equation*} {{I}^{(n)}}\left( c,k \right)=\sum\limits_{i=0}^{\infty }{\left\{ \frac{{{\left( -1 \right)}^{n-1+i}}}{\left( n-1 \right)!}\frac{{{d}^{(n-1)}}}{dk}\left[ \frac{1}{\prod\limits_{j=0}^{i}{\left( k+1+j \right)}} \right]{{I}^{(0)}}\left( c,k+i \right) \right\}}.\tag{2.7} \end{equation*}


[Proof for (2.7)]

Also, we can reduce h_{k-r}^{c} to

\begin{equation} h_{k-r}^{c}={{c}^{-k-1}}{\exp(-c)}{{I}^{\left( 0 \right)}}\left( c,k \right)+{{c}^{-k-1}}{\exp(-c)}r\sum\limits_{i=0}^{\infty }{\frac{{\left( -1 \right)}^{i}{{I}^{\left( 0 \right)}}\left( c,k+i \right)}{\prod\limits_{j=0}^{i}{\left( k+1+j-r \right)}}}. \tag{2.8} \end{equation}


[Proof for (2.8)]

As mentioned earlier, we are interested in the gamma integral where the base parameter s is bounded within the interval \left( 0,1 \right). Applying the "h" factorization as shown in (2.2), we derive -1<k-r< 0 and k=0, given 0< r<1. Thus,

\begin{align} h_{-r}^{c}&={{c}^{-1}}{\exp(-c)}\left\{ \left( {\exp(c)}-1 \right)+r\left( \frac{{\exp(c)}-1}{0!} \right)\left( \frac{1}{1-r} \right) \right. \tag{2.9} \\ \notag & +r\left( \frac{c \cdot {\exp(c)}-{\exp(c)}+1}{1!} \right)\left( \frac{1}{2-r}-\frac{1}{1-r} \right) \\ \notag & \left. +r\left( \frac{{{c}^{2}}\cdot {\exp(c)}-2c \cdot {\exp(c)}+2! \cdot {\exp(c)}-2!}{2!} \right)\left( \frac{1}{3-r}-\frac{2}{2-r}+\frac{1}{1-r} \right)+\cdots \right\}. \end{align}


[Proof for (2.9)]

Since \Gamma \left( -r+1 \right)={{c}^{-r+1}}{\exp(-c)}h_{-r}^{-c}{{\Bigr|}_{c\to \infty }}, we can respecify (2.9) and replace the power-term parameter c with -c


\begin{equation} h_{-r}^{-c}={{c}^{-1}}{\exp(c)}\left[1-{\exp(-c)} \right]+{{c}^{-1}}{\exp(c)}r\sum\limits_{i=0}^{n }{{{w}_{i}}{{\beta }_{i}}}, \tag{2.10} \end{equation}



where {{w}_{i}}=1-\sum\limits_{j=0}^{i}{\frac{{{c}^{j}}{\exp(-c)}}{j!}}, {{\beta }_{i}}=\frac{i!}{\prod\limits_{j=0}^{i}{\left( j+1-r \right)}}, and n\to \infty .
[Proof for (2.10)]

As demonstrated below, we can reduce h_{-r}^{-c} to its simplest form, which directly links to the gamma function of \Gamma \left( -r \right)

\begin{equation} h_{-r}^{-c}=-{{c}^{-1}}-r{{c}^{r-1}}{\exp(c)}\Gamma \left( -r \right). \tag{2.11} \end{equation}


[Proof for (2.11)]

Therefore, "h" explains the recursive rule of the gamma function

\begin{align*} \Gamma \left( -r+1 \right)&={{c}^{-r+1}}{\exp(-c)}h_{-r}^{-c}{{|}_{c\to \infty }} \tag{2.12} \\ \notag & =-r\Gamma \left( -r \right) \notag. \end{align*}


[Proof for (2.12)]

This finding not only defines the algebraic meaning of the "h" function by connecting to the gamma function, but also generalizes the factorial function to non-integer cases. As is evident in (2.12), we can explicitly define \Gamma \left( s \right) as a function of h_{s-1}^{-c}{{|}_{c\to \infty }}, where 0<s<1. Hence, all gamma functions defined in (2.3) can be expressed as an "h" function for all real arguments, except for the case of non-positive integers.

 

 

 

 

 

 

 

 

 

Download [full paper] [supplementary materials] [.m files] [technical note]