Pentatope Numbers

Today I turned 26176 days old and one of the properties of this number, as listed in the Online Encyclopaedia of Integer Sequences (OEIS), is that it's a member of OEIS A118411:

A118411

Numerator of sum of reciprocals of first n pentatope numbers A000332.

 The sequence runs:

1, 6, 19, 136, 83, 119, 656, 73, 190, 121, 1816, 559, 679, 815, 3872, 1139, 886, 513, 2360, 2023, 2299, 2599, 11696, 3275, 7306, 1353, 5992, 1653, 5455, 5983, 26176, ...

I didn't know what pentatope numbers were so I needed to find out. Firstly, this type of number is a particular example of a more general type of number, the polytope number. It can exist in any number of dimensions and so we can speak of an \(n\)-polytope where \(n\) represents the number of dimensions. However, we get into deep water quickly if we wax too general so I'll bring it back to specifics for the moment by looking at Pascal's triangle (Figure 1).

Figure 1: Derivation of pentatope numbers from
    a left-justified Pascal's triangle. Source.

So a pentatope number could be defined as any number in the fifth cell of any row of Pascal's triangle starting with the 5-term row 1 4 6 4 1. The terms are given by OEIS A000332:

1, 5, 15, 35, 70, 126, 210, 330, 495, 715, 1001, 1365, 1820, 2380, 3060, 3876, 4845, 5985, 7315, 8855, 10626, 12650, 14950, 17550, 20475, 23751, 27405, 31465, 35960, ...

Pascal's triangle lists the binomial coefficients and the pentatope numbers (let's refer to them as \(P_n\)), can be expressed in terms of these coefficients, specifically:$$P_n=\binom{n+3}{4} \text{ where }n \geq 1$$This reads as the number of ways to choose \(n+3\) things four at a time. An alternative representation involves the use of the rising factorial, so that we can write:$$P_n=\frac {n^{ \overline 4}}{4!}=\frac{n(n+1)(n+2)(n-3)}{24} \text { where }n \geq 1$$The infinite sum of the reciprocals of all pentatope numbers is \(\frac{4}{3} \) and this takes us back to OEIS A118411, the sequence that started this post. In other words:$$ \sum^ {\infty}_{n=1} \frac{4!}{n(n+1)(n+2)(n+3)}=\frac{4}{3}$$This sequence only shows the numerators, so let's write some SageMath code to generate the actual fractions (permalink). See Figure 2.

Figure 2

The result is:

(1, 1), (6, 5), (19, 15), (136, 105), (55, 42), (83, 63), (119, 90), (656, 495), (73, 55), (190, 143), (121, 91), (1816, 1365), (559, 420), (679, 510), (815, 612), (3872, 2907), (1139, 855), (886, 665), (513, 385), (2360, 1771), (2023, 1518), (2299, 1725), (2599, 1950), (11696, 8775), (3275, 2457), (7306, 5481), (1353, 1015), (5992, 4495), (1653, 1240), (5455, 4092), (5983, 4488), (26176, 19635)

Now \(\frac{4}{3}=1. \dot{3} \) and even \( \frac{6}{5}=1.2 \) is getting close to this. If we look at the final fraction in the above list, we see that: $$\frac{26176}{19635}=1.33312961548256$$ Clearly the fractions are getting closer to \(1. \dot{3}\). The generating function for the pentatope numbers is:$$ \frac{x}{(1-x)^5}=x+5x^2+15x^3+35x^4+...$$Referring back to Figure 1, it can be seen that the pentatope numbers are the cumulative sums of the tetrahedral number, just as the tetrahedral numbers are the cumulative sums of the triangular numbers and the triangular numbers are the cumulative sums of the natural numbers. 

To quote from this source:

"Pentatope" is a recent term. Regarding the fifth row, Pascal wrote that ... since there are no fixed names for them, they might be called triangulo-triangular numbers. Pentatope numbers exists in the 4D space and describe the number of vertices in a configuration of 3D tetrahedrons joined at the faces.

This is a big topic and a lot more could be said of about pentatopes and polytopes but that will do for now as we are focusing on the pentatope numbers. 

Permutahedron

On Quanta, I came across an interesting interview with a mathematician, Federico Ardila, who speaks about the connection between combinatorics and geometry:

When you look at the geometric side of things, suppose, for example, you want to study the permutations (the ways of rearranging a collection of objects). It’s pretty well known that if you have n objects, the number of ways of putting them in a row is n factorial (the product n(n-1)(n-2)…1). So it’s not a very interesting problem to count how many ways there are. But what is their inherent structure?  

The three-dimensional permutahedron, a geometric depiction of the ways to rearrange the numbers 1, 2, 3 and 4. Two permutations are connected by an edge if one can be transformed into the other by swapping two consecutive numbers.     

If you look at when two permutations are related to each other by just swapping two elements, then you start understanding not only how many there are but how are they related to each other. And then, when you say, “OK, let’s take all the permutations, and put an edge between two of them if they’re a swap away,” then you find that you get this beautiful shape that’s a polytope (a geometric object with flat sides). I think it’s completely surprising initially that the inherent relations between permutations are captured in this beautiful polytope called a permutahedron. So all of a sudden you have this geometric model, and you can use tools from polytope theory to try to say new things about permutations. And that polytope has existed for a long time and is very well understood.

The three-dimensional permutahedron, a geometric depiction of the ways to rearrange the numbers 1, 2, 3 and 4. Two permutations are connected by an edge if one can be transformed into the other by swapping two consecutive numbers

This particular structure is in fact a truncated octahedron with fourteen faces comprised of six squares and eight hexagons. There are 24 vertices and 36 edges. Whereas the octahedron is a Platonic solid, the truncated version is an Archimedean solid defined as follows:

In geometry, an Archimedean solid is one of the 13 solids first enumerated by Archimedes. They are the semi-regular convex polyhedra composed of regular polygons meeting in identical vertices, excluding the 5 Platonic solids (which are composed of only one type of polygon) and excluding the prisms and antiprisms. They differ from the Johnson solids, whose regular polygonal faces do not meet in identical vertices. Source.

An example of a Johnson solid would be a triangular bipyramid or dipyramid, a type of hexahedron formed by joining two identical tetrahedra together, the face on one being exactly against a face of the other. The truncated octahedron ends up looking like this:

Interestingly the above shape can, like the cube, tesselate or pack 3-dimensional space. It's by no means obvious but, quite amazingly, this is what happens:

In Mathematics, a permutahedron is an (n−1)-dimensional polytope embedded in an n-dimensional space, the vertices of which are formed by permuting the coordinates of the vector (1, 2, 3, ..., n). Thus the permutahedron of order 4 exists in 3 dimensions, order 3 in two dimensions and order 2 in 1 dimension. A polytope is defined as:

... a geometric object with "flat" sides. It is a generalisation in any number of dimensions of the three-dimensional polyhedron. Polytopes may exist in any general number of dimensions n as an n-dimensional polytope or n-polytope. Flat sides mean that the sides of a (k+1)-polytope consist of k-polytopes that may have (k-1)-polytopes in common. For example, a two-dimensional polygon is a 2-polytope and a three-dimensional polyhedron is a 3-polytope. Source.

All this takes us into complicated areas of mathematics but what's important here is simply to note the link has been forged between two apparently disparate mathematical topics: combinatorics and geometry.