Intersections between this infinite power tower and a multifactorial

In summary, multifactorial notation can be confusing for those who are unfamiliar with it. Common mistakes include confusing the notation with factorials and not understanding its definition for different values of n. To better understand multifactorials, a function called f_k(x) is defined, which is real and continuous for all values of x. Another important concept related to multifactorials is tetration, which can be expressed using Knuth's up-arrow notation. One interesting relationship that arises from tetration is the function h(z), which is defined using the Lambert W function and only converges for values of z in the domain [e^-e, e^1/e]. To find the smallest integer value of k for which f_k(x) and h
  • #1
Saracen Rue
150
10
TL;DR Summary
Find the smallest integer value for ##k## for which ##x!^{k}## has the largest number of intersections with ##{}^{\infty}(!x)##, and state the values of ##x## and ##y## at each intersection point.
For those unaware of multifactorial notation, it should be noted that there are some common mistakes made when first being introduced to the notation. For example, ##n! \neq (n!)!## and ##n! \neq (n!)! \neq (n!)! \neq ((n!)!)!##. Just to make sure we're all up to speed, here's a quick run down of how multifactorials are defined for integer values of ##n##: $$n!=n(n-1)(n-2)(n-3)...(n-a)|\text{ } (n-a) > 0$$ $$n!=n(n-2)(n-4)(n-6)...(n-a)|\text{ } 2 \geq (n-a) > 0$$ $$n!=n(n-3)(n-6)(n-9)...(n-a)|\text{ } 3 \geq (n-a) > 0$$ $$\large \text{For a more generalized definition:}$$ $$n!^{k}= \left. \begin{cases}
1 & n=0\\
n & 1\leq n\leq k \\
n(n-k)(n-2k)(n-3k)...(n-a) & n>k
\end{cases}\right| k \geq (n-a) > 0$$ Continuing on from this, the next step is to define a function, ##f_k(x)=x!^{k}##, such that ##f_k## is real and continuous for all values of ##x##, which can be accomplished like so:
$$x!^{k}= \left. k^{\frac{x}{k}} \cdot \Gamma(1+\frac{x}{k}) {\prod_{j=1}^{k}}\left(\left(\frac{j \cdot k^{\frac{-j}{k}} }{\Gamma(1+ \frac{j}{k})}\right)^{\displaystyle \frac{1}{k} \sum_{n=1}^{k}\cos\left(\frac{2\pi n\left(x-j\right)}{k}\right)}\right)\right| x \in \mathbb{R}, k \in \mathbb{Z}^+$$ $$\large { \text{Where } !^{k} \text{ represents }k\text{ number of factorial symbols after }} x$$ Now for the second function in this question we need to talk about tetration and subfactorials. Focusing on the latter for now, let's look at the definition for a subfactorial (also known as derangement): $$!n=\text{subfactorial}(n)=\frac{\Gamma(n+1, -1)}{e}$$ $$\large \text{Where } \Gamma(a, x) \text{ is the incomplete gamma function}$$ This leaves us with the final topic of tetration. Tetration is a very useful tool and can be seen as another way of expressing Knuth's up-arrow notation. Consider the following: $$^{n}a=a\uparrow\uparrow n = \Large\underbrace{a^{a^{a^{.^{.^{.^a}}}}}}_{n \text{ times}}$$ Note: the notation can be kind of tricky here; "##n## times" means that the total number of '##a##'s present is equal to ##n##. For example: $$\Large ^{4}3={3^{3^{3^{3}}}}=10^{10^{12.56090264130034}}$$ When this is then applied to create an infinitely tall power tower we end up with this interesting relationship: $$\large ^{\infty}z=z^{z^{z^{.^{.^{.^z}}}}}=h(z)=\frac{-W(-\ln(z))}{\ln(z)}$$ $$\large \text{Where } W(z) \text{ is the Lambert W function}$$ It should be duly noted that ##h(z)## will converge if and only if ##z## falls in the domain ##[e^{-e}, e^{\frac{1} {e}}]##. $$\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_$$ Now, with the definitions out of the way and hopefully everyone is on the same page, we can finally address the question; $$\large \text{Find the smallest integer value of } k \text{ for which}$$ $$\large f_k(x)=x!^{k} \text{ and }h(!x)=\frac{-W(-\ln(!x))}{\ln(!x)}$$ $$\large \text{have the most points of intersection for } x \geq 0$$ $$\large \text{and evaluate the coordinates for each point, correct to }3 \text{ decimal places}$$
 
Physics news on Phys.org
  • #2
Regarding this comment:

Saracen Rue said:
When this is then applied to create an infinitely tall power tower we end up with this interesting relationship: $$\large ^{\infty}z=z^{z^{z^{.^{.^{.^z}}}}}=h(z)=\frac{-W(-\ln(z))}{\ln(z)}$$ $$\large \text{Where } W(z) \text{ is the Lambert W function}$$ It should be duly noted that ##h(z)## will converge if and only if ##z## falls in the domain ##[e^{-e}, e^{\frac{1} {e}}]##.
I believe it's more accurate to write:
Let:
$$y=\large^{\infty} z=z^{z^{z^{.^{.^{.}}}}}$$
and
$$D=\{z\in\mathbb{C}: y(z)\;\text{converges}\}$$
Then ##y=z^y## for ##z\in D## and we may write:
$$
y(z)=\frac{-\text{W}(-\log(z))}{\log(z)}; \;\; z\in D
$$
but the function
$$h(z)=-\frac{\text{W}(-\log(z))}{\log(z)}$$
is analytic (and multivalued) everywhere except at it's singular points.

Regarding this expression:

$$\large f_k(x)=x!^{k} \text{ and }h(!x)=\frac{-W(-\ln(!x))}{\ln(!x)}$$

If we plot the trajectory of ## !x## then when ##x=1## it's passing through the logarithmic branch point. Perhaps, though you could perturb the path around the origin and remain in analytic continuity on the same principal-valued sheet or even drop to the next sheet perturbing it in the opposite fashion and still remain analytically continuous. Not sure though how that would affect matters.
powertowerpath.jpg
 
Last edited by a moderator:
Back
Top