62
$\begingroup$

Introduction:

We know that:

$$\sum_{x=0}^\infty \frac{1}{x!}=e$$

But what if we replaced $x!$ with $!x$ also called the subfactorial function also called the $x$th derangement number? This interestingly is just a multiple of $e$ and an Incomplete Gamma function based sum. This will get us a new number as the $n=0$ and 1 terms diverge as a result of the reciprocal. I also use the Generalized Exponential Integral function and the Round function. The OEIS entry for the constant is A281682:

\begin{align*} S &=\sum_{x=2}^\infty \frac{1}{!x} =\sum_{n=2}^\infty \frac{1}{\operatorname{Round}\bigl(\frac{x!}{e}\bigr)} =e\sum_{x=2}^\infty\frac{1}{Γ(x+1,-1)}= \\ &= e\sum_{n=3}^\infty \frac 1{Γ(x,-1)}=\sum_{X=0}^\infty \sum_{x=2}^\infty\frac{1}{Γ(X+1)Γ(x+1,-1)} =-e\sum_{x=2}^\infty\frac{(-1)^x}{E_{-x}(-1)}= -e\sum_{x=-\infty}^{-2}\frac{(-1)^x}{E_x(-1)} =1.63822707… \end{align*}

Possible Abel-Plana formula Application:

We can also use the Abel-Plana formula, and the alternate series version, to find an integral representation of the sum. You can also use other representations of the summand, but this integral is probably hard to work with.

Note the Abel-Plana formula may not work with the constant:

\begin{align*} S &=\sum_{x=0}^\infty\frac{1}{!(x+2)} =\frac{1}{2} + \int_0^\infty \frac{dx}{!(x+2)} + i\int_0^\infty\frac{\frac{1}{!(2+ix)}-\frac{1}{!(2-ix)}}{e^{2\pi x}-1} \, dx \\ \implies -\frac{S}{e} &=\sum_{x=0}^\infty\frac{(-1)^x}{E_{-x-2}(-1)}=-\frac{1}{2e} + \frac i2\int_0^\infty \left[\frac{1}{E_{-ix-2}(-1)}-\frac{1}{E_{ix-2}(-1)}\right] \operatorname{csch}(\pi x) \, dx \end{align*}

See this nice closed form result of

$$\sum_{x=-\infty}^0 \text {Im}(!x)=-\frac{\pi}{e^2}$$

I do not think this simple looking problem has been posted so far.

The sum does not need to be in closed form.

You also can rewrite it in terms of a better sum. I am more looking for an evaluation or manipulation of the sum. Please correct any mistakes and give me feedback!

A Mittag-Leffler Insight:

Because $$\sum_{x=2}^\infty \frac{1}{!x} =e\sum_{x= 2}^\infty\frac{1}{Γ(x+1,-1)}= e\sum_{n=3}^\infty\frac 1{Γ(x,-1)} $$

one may notice the relation to the Mittag-Leffler function:

$$\text E_{a,b}(x)=\sum_{n=0}^\infty \frac{x^n}{Γ(ax+b)}$$

The only problem is if there existed a function for the incomplete gamma function analogue of the Mittag-Leffler function. Maybe one can find this function or use the already known one?

Another Integral Representation:

It can be shown that the following is true using @Jack Barber’s method in

$$\sum_\Bbb N \text{erfc}(x)$$

Here is an integral representation using the linearity of the Floor function and the Meijer G function:

$$\sum_2^\infty \frac{1}{!x}=-\int_2^\infty \lfloor x-1\rfloor \frac{d}{dx} \frac{1}{!x}dx=\int_2^\infty \frac{d}{dx} \frac{1}{!x} dx-\int_2^\infty\lfloor x\rfloor \frac{d}{dx} \frac{1}{!x} dx=\frac1{!\infty}-\frac1{!2}-\int_2^\infty \lfloor x \rfloor \frac{d}{dx} \frac{1}{!x} =-1-\int_2^\infty \lfloor x \rfloor \frac{d}{dx} \frac{1}{!x}dx= -1-\int_2^\infty \lfloor x \rfloor\left(-\frac{\text G_{2,3}^{3,0}\big(-1\left|_{0,0,x+1}^{\ \ \ \ 1,1}\right)}{e(!x)^2}-\frac{i\pi}{!x}\right)dx =\frac1e\int_2^\infty \frac{\lfloor x\rfloor\text G_{2,3}^{3,0}\left(-1\big|_{0,0,x+1}^{\ \ \ \ 1,1}\right)}{(!x)^2} dx+i\pi\int_2^\infty\frac{\lfloor x\rfloor}{!x}dx-1$$

The Meijer G function is hard to use, but you can come up with many more integral representations using alternate forms of the floor function; there is even one in terms of elementary functions.

A Manageable Series Expansion:

Converting the Round function to elementary functions, we have this form:

$$e\sum_{m=0}^\infty \sum_{n=2}^\infty \frac{\left(\frac e\pi\frac{\tan^{-1}\left(\tan\left(\frac\pi en!\right) \right)}{n!}\right)^m}{n!}$$

After a bit more work, we remove $!n$ from the denominator replacing it with gamma regularized $Q(a,z)$:

$$S=\sum_{m=0}^\infty\sum_{k=0}^n(-1)^k e^{k+1}\binom mk\sum_{n=2}^\infty\frac{(!n)^k}{n!^{k+1}}=\boxed{e\sum_{m,k=0}^\infty\binom mk(-1)^k\sum_{n=2}^\infty\frac{Q^k(n+1,-1)}{n!}}$$

Shown here.

Conclusion:

A closed form is optional, but good alternative representations for $S$ would also work. Please do not make up any new function. Alternatively, what is $\sum\limits_{n=2}^\infty\frac{Q^k(n+1,-1)}{n!}$?

$\endgroup$
29
  • 10
    $\begingroup$ $\sum\frac1{!x}$ converges at the same rate as $\sum\frac1{x!}$, which is already faster than exponential; asking for even better convergence seems very ambitious! $\endgroup$ Aug 6, 2021 at 20:30
  • 2
    $\begingroup$ @AlBrown Please see this definition from Wolfram Functions. From -1 to infinity diverges. $\endgroup$ Aug 6, 2021 at 21:21
  • 2
    $\begingroup$ @AlBrown Does this help? I would have done a wolfram alpha link, but it refuses to work, so here is a results photo. Notice the end result is the same. $\endgroup$ Aug 6, 2021 at 21:29
  • 2
    $\begingroup$ Plus i saw what you added. I meant integral from -1, but im good $\endgroup$
    – Al Brown
    Aug 6, 2021 at 22:12
  • 3
    $\begingroup$ @Tyma I wasn't (only) talking about practical applications of this "derangement constant". I don't think it has much application in pure mathematics either, (but I'm certainly no expert on this topic), whereas constants like pi & e are relevant to many areas in pure mathematics. I agree that it is interesting to see alternate forms for this constant, even if they don't necessarily lead to more efficient ways to calculate it. $\endgroup$
    – PM 2Ring
    Sep 28, 2021 at 5:49

0

You must log in to answer this question.