0
$\begingroup$

Is it possible to further simplify the expression

$$\sum\limits_{n = 0}^{\left\lfloor {\frac{M}{2}} \right\rfloor } {\frac{{M!}}{{{2^n} \times \left( {n!} \right) \times \left( {M - 2n} \right)!}}},$$

where $M$ is a positive integer?

$\endgroup$
3
  • 1
    $\begingroup$ it depends ... if you think that the holonomic sequence, or the hypergeometric function are a simplification $\endgroup$ Sep 29 '20 at 14:23
  • 2
    $\begingroup$ Notice that this is the number of involutions on permutations of $M$ elements. $\endgroup$
    – Phicar
    Sep 29 '20 at 14:33
  • $\begingroup$ General strategy ... calculate the first few values & then look up the sequence in OEIS. $\endgroup$ Sep 29 '20 at 20:31
1
$\begingroup$

Making the problem more general $$S_M=\sum\limits_{n = 0}^{\left\lfloor {\frac{M}{2}} \right\rfloor } {\frac{{M!}}{{ {n!} \, \left( {M - 2n} \right)!}}}x^n$$ $$S_{2m}=(-1)^m (4x)^m \,U\left(-m,\frac{1}{2},-\frac{1}{4 x}\right)$$ $$S_{2m+1}=(-1)^m (4x)^m \,U\left(-m,\frac{3}{2},-\frac{1}{4 x}\right)$$ where appears the confluent hypergeometric function.

If $x=\frac 12$, these generate the sequences $$\{2,10,76,764,9496,140152,2390480,46206736,997313824,23758664096\}$$ which correspond to $$S_{2m}=2^n n!\, L_n^{-\frac{1}{2}}\left(-\frac{1}{2}\right)$$ and $$\{4,26,232,2620,35696,568504,10349536,211799312,4809701440,119952692896\}$$ which correspond to $$S_{2m+1}=2^n n!\, L_n^{\frac{1}{2}}\left(-\frac{1}{2}\right)$$ where appear the generalized Laguerre polynomials.

Have a look at $OEIS$ for very interesting informations.

$\endgroup$
2
  • $\begingroup$ Could you kindly confirm that $U\left( {} \right)$ represent the Tricomi's (confluent hypergeometric) function U(a, b, z) (introduced by Francesco Tricomi (1947)), which is a solution to Kummer's equation. Or also known as the confluent hypergeometric function of the second kind ? and $L_n^\alpha \left( x \right) = \frac{{{{\left( { - 1} \right)}^n}}}{{n!}}U\left( { - n,\alpha + 1.x} \right)$ $\endgroup$ Sep 30 '20 at 11:07
  • 1
    $\begingroup$ @TuongNguyenMinh. Yes, it is $\endgroup$ Sep 30 '20 at 11:25

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.