My Favorite Sequences: A261865

This is the first installment in a new series, “My Favorite Sequences”. In this series, I will write about sequences from the On-Line Encyclopedia of Integer Sequences that I’ve authored or spent a lot of time thinking about.

I’ve been contributing to the On-Line Encyclopedia of Integer Sequences since I was an undergraduate. In December 2013, I submitted sequence A233421 based on problem A2 from the 2013 Putnam Exam—which is itself based on “Ron Graham’s Sequence” (A006255)—a surprising bijection from the natural numbers to the non-primes. As of today, I’ve authored over 475 sequences based on puzzles that I’ve heard about and problems that I’ve dreamed up.

A261865: Multiples of square roots

(This problem is closely related to Problem 13 in my Open Problems Collection.)

In September 2015, I submitted sequence A261865:

\(A261865(n)\) is the least integer \(k\) such that some multiple of \(\sqrt k\) falls in the interval \((n, n+1)\).

An illustration of the first dozen terms of A261865

For example, \(A261865(3) = 3\) because there is no multiple of \(\sqrt 1\) in \((3,4)\) (since \(3 \sqrt{1} \leq 3\) and \(4 \sqrt{1} \geq 4\)); there is no multiple of \(\sqrt{2}\) in \((3,4)\) (since \(2 \sqrt{2} \leq 3\) and \(3 \sqrt 2 \geq 4\)); but there is a multiple of \(\sqrt 3\) in \((3,4)\), namely \(2\sqrt 3\).

As indicated in the picture, the sequence begins $$\color{blue}{ 2,2,3,2,2},\color{red}{3},\color{blue}{2,2,2},\color{red}{3},\color{blue}{2,2},\color{red}{3},\color{blue}{2,2,2},\color{red}{3},\color{blue}{2,2},\color{red}{3},\color{blue}{2,2},\color{magenta}{7},\dots.$$

A scatterplot of \(A261865(n)\). Notice the records at \(A261865(184)=38\) and \(A261865(8091)=43\).

A conjecture about density

As the example illustrates, \(1\) does not appear in the sequence. And almost by definition, asymptotically \(1/\sqrt 2\) of the values are \(2\)s.

Let’s denote the asymptotic density of terms that are equal to \(n\) by \(d_n\). It’s easy to check that \(d_1 = 0\), (because multiples of \(\sqrt 1\) are never between any integers) and \(d_2 = 1/\sqrt 2\), because multiples of \(\sqrt 2\) are always inserted. I conjecture in Problem 13 of my Open Problem Collection that $$a_n = \begin{cases}\displaystyle\frac{1}{\sqrt n}\left(1 – \sum_{i=1}^{n-1} a_i\right) & n \text{ is squarefree}\\[5mm] 0 & \text{otherwise}\end{cases}$$

If this conjecture is true, then the following table gives approximate densities.

\(i\)\(d_i\)
\(1\)\(d_1 = 0\%\)
\(2\)\(d_2 = 70.7\%\)
\(3\)\(d_3 = 16.9\%\)
\(4\)\(d_4 = 0\%\)
\(5\)\(d_5 = 5.54\%\)
\(6\)\(d_6 = 2.79\%\)
\(7\)\(d_7 = 1.53\%\)
\(10\)\(d_{10} = 0.797\%\)
\(11\)\(d_{11} = 0.519\% \)
\(399\)\(d_{399} = 3.53 \times 10^{-11} \%\)

This was computed with the Mathematica code:

d[i_] := (d[i] = If[
  SquareFreeQ[i], 
  N[(1 - Sum[d[j], {j, 2, i - 1}])/Sqrt[i], 50], 
  0
])

Finding Large Values

I’m interested in values of \(n\) such that \(A261865(n)\) is large, and I reckon that there are clever ways to construct these, perhaps by looking at some Diophantine approximations of \(\sqrt{2}, \sqrt{3}, \sqrt{5}, \sqrt{6}, \dots\). In February, I posted a challenge on Code Golf Stack Exchange to have folks compete in writing programs that can quickly find large values of \(A261865(n)\).

Impressively, Noodle9’s C++ program won the challenge. In under a minute, this program found that the input \(n=1001313673399\) makes \(A261865\) particularly large: \(A261865(1001313673399) = 399\). Within the time limit, no other programs could find a value of \(n\) that makes \(A261865(n)\) larger.

\(n\)Order of magnitude\(A261865(n)\)Time
1 \(1 \times 10^{0}\)2(0s)
3 \(3 \times 10^{0}\)3 (0s)
23 \(2.3 \times 10^{1}\)7 (0s)
30 \(3.0 \times 10^{1}\)15 (0s)
184 \(1.84 \times 10^{2}\)38 (0s)
8091 \(8.091 \times 10^{3}\)43 (0s)
16060 \(1.606 \times 10^{4}\)46 (0s)
16907 \(1.691 \times 10^{4}\)58 (0s)
20993 \(2.099 \times 10^{4}\)61 (0s)
26286 \(2.629 \times 10^{4}\)97 (0s)
130375 \(1.304 \times 10^{5}\)118 (0s)
169819 \(1.698 \times 10^{5}\)127 (0s)
2135662 \(2.136 \times 10^{6}\)130 (0s)
2345213 \(2.345 \times 10^{6}\)187 (0s)
46272966 \(4.627 \times 10^{7}\)193 (1s)
222125822 \(2.221 \times 10^{8}\)210 (5.2s)
237941698 \(2.379 \times 10^{8}\)217 (5.7s)
257240414 \(2.572 \times 10^{8}\)227 (6.2s)
1205703469 \(1.206 \times 10^{9}\)267 (31s)
1558293414 \(1.558 \times 10^{9}\)299 (41.8s)
4641799364 \(4.642 \times 10^{9}\)303 (2.1m)
6600656102 \(6.601 \times 10^{9}\)323 (3m)
11145613453 \(1.115 \times 10^{10}\)335 (5.2m)
20641456345 \(2.064 \times 10^{10}\)354 (9.8m)
47964301877 \(4.796 \times 10^{10}\)358 (22.9m)
105991039757 \(1.06 \times 10^{11}\)385 (52m)
119034690206 \(1.19 \times 10^{11}\)397 (59.1m)
734197670865 \(7.342 \times 10^{11}\)455 (6.4h)
931392113477 \(9.314 \times 10^{11}\)501 (8.4h)
1560674332481 \(1.561 \times 10^{12}\)505 (14.2h)
A table of record values as computed by Code Golf Stack Exchange user Neil. The first 16 values agree with Jon E. Schoenfield’s computations that were added to the OEIS in September 2015

Related Ideas

Sequence \(A327953(n)\) counts the number of positive integers \(k\) such that there is some integer \(\alpha^{(n)}_k > 2\) where \(\alpha^{(n)}_k\sqrt{k} \in (n, n+1)\). It appears to grow roughly linearly like \(A327953(n) \sim 1.3n\), but I don’t know how to prove this.

  • Take any function \(f\colon\mathbb N \rightarrow \mathbb R\) that is positive, has positive first derivative, and has negative second derivative. Then, what is the least \(k\) such that some multiple of \(f(k)\) is in \((n,n+1)\)?
  • For example, what is the least integer \(k \geq 3\) such that there is a multiple of \(\ln(k)\) in \((n, n+1)\)?
  • What is the least \(k \in \mathbb N\) such that there exists \(m \in \mathbb N\) with \(k2^{1/m} \in (n,n+1)\)?
  • What is the least \(m \in \mathbb N\) such that there exists \(k \in \mathbb N\) with \(k2^{1/m} \in (n,n+1)\)?
  • A343205 is the auxiliary sequence that gives the value \(m\) such that \(m\sqrt{A261865(n)} \in (n, n+1)\). Does this sequence have an infinite limit inferior?
Scatterplot of A343205, generated in Mathematica. If the main conjecture is true, then this is not bounded below by \(\alpha n\) for any positive value of \(\alpha\).

If you can answer any of these questions, or if you spend time thinking about this, please let me know on Twitter, @PeterKagey!

Richard Guy’s Partition Sequence

Neil Sloane is the founder of the On-Line Encyclopedia of Integer Sequences (OEIS). Every year or so, he gives a talk at Rutgers in which he discusses some of his favorite recent sequences. In 2017, he spent some time talking about a 1971 letter that he got from Richard Guy, and some questions that went along with it. In response to the talk, I investigated the letter and was able to sort out Richard’s 45-year-old idea, and correct and compute some more terms of his sequence.

Richard Guy and his sequences

Richard Guy was a remarkable mathematician who lived to the remarkable age of 103 years, 5 months, and 9 days! His life was filled with friendships and collaborations with many of the giants of recreational math: folks like John Conway, Paul Erdős, Martin Gardner, Donald Knuth, and Neil Sloane. But what I love most about Richard is how much joy and wonder he found in math. (Well, that and his life-long infatuation with his wife Louise.)

Richard guy mountaineering at age 98 with a photo of his late wife, Louise.

[I’m] an amateur [mathematician], I mean I’m not a professional mathematician. I’m an amateur in the more genuine sense of the word in that I love mathematics and I would like everybody in the world to like mathematics.

Richard Guy in Fascinating Mathematical People: Interviews and Memoirs

Richard’s letter to Neil

In January 2017, Neil Sloane gave a talk at Doron Zeilberger’s Experimental Mathematics Seminar, and about six minutes in, Neil discusses a letter that Richard sent to him at Cornell—which was the forwarded to Bell Labs—in June 1971.

Richard Guy’s 1971 letter to Neil Sloane.

When I was working on the book, the 1973 Handbook of Integer Sequences, I would get letters from Richard Guy from all over the world. As he traveled around, he would collect sequences and send them to me.

Neil Sloane, Rutgers Experimental Mathematics Seminar

At 11:30, Neil discusses “sequence I” from Richard’s letter, which he added to the OEIS as sequence A279197:

Number of self-conjugate inseparable solutions of \(X + Y = 2Z\) (integer, disjoint triples from \(\{1,2,3,\dots,3n\}\)).

Neil mentioned in the seminar that he didn’t really know exactly what the definition meant. With some sleuthing and programming, I was able to make sense of the definition, write a Haskell program, correct the 7th term, and extend the sequence by a bit. The solutions for \(A279197(1)\) through \(A279197(10)\) are listed in a file I uploaded to the OEIS, and Fausto A. C. Cariboni was able to extend the sequence even further, submitting terms \(A279197(11)\)–\(A279197(17)\).

How the sequence works.

The idea here is to partition \(\{1,2,3,\dots,3n\}\) into length-3 arithmetic progressions, \(\bigl\{\{X_i,Z_i,Y_i\}\bigr\}_{i=1}^{n}\). And in particular, we want them to be inseparable and self-conjugate.

An inseparable partition is one whose “smallest” subsets are not a solution for a smaller case. For example, if \(n=3\), then the partition \[\bigl\{ \{1,3,5\}, \{2,4,6\}, \{7,8,9\} \bigr\}\] is separable, because if the subset \(\bigl\{ \{1,3,5\}, \{2,4,6\} \bigr\}\) is a solution to the \(n=2\) case.

A self-conjugate partition is one in which swapping each \(i\) with each \(3n+1-i\) gets back to what we started with. For example, \(\bigl\{\{1,3,5\}, \{2,4,6\}\bigr\}\) is self-congugate, because if we replace the \(1\) with a \(6\) and the \(2\) with a \(5\), and the \(i\) with a \(7-i\), then we get the same set: \(\bigl\{\{6,4,2\}, \{5,3,1\} \bigr\}\)

(1,3,5),  (2,7,12), (4,9,14),  (6,8,10),  (11,13,15)
(1,3,5),  (2,8,14), (4,7,10),  (6,9,12),  (11,13,15)
(1,5,9),  (2,3,4),  (6,8,10),  (7,11,15), (12,13,14)
(1,5,9),  (2,4,6),  (3,8,13),  (7,11,15), (10,12,14)
(1,6,11), (2,3,4),  (5,10,15), (7,8,9),   (12,13,14)
(1,6,11), (2,7,12), (3,8,13),  (4,9,14),  (5,10,15)
(1,7,13), (2,4,6),  (3,9,15),  (5,8,11),  (10,12,14)
(1,7,13), (2,8,14), (3,9,15),  (4,5,6),   (10,11,12)
(1,8,15), (2,3,4),  (5,6,7),   (9,10,11), (12,13,14)
(1,8,15), (2,4,6),  (3,5,7),   (9,11,13), (10,12,14)
(1,8,15), (2,4,6),  (3,7,11),  (5,9,13),  (10,12,14)
Each line shows one of the \(A279197(5) = 11\) inseparable, self-conjugate partitions of \(\{1,2,\dots,15\}\).

Generalizing Richard Guy’s idea

Of course, it’s natural to wonder about the separable solutions, or what happens if the self-conjugate restriction is dropped. In exploring these cases, I found four cases already in the OEIS, and I computed five more: A282615A282619.

SeparableInseparableEither
Self-conjugateA282615A279197 A282616
Non-self-conjugateA282618A282617 A282619
EitherA279199A202705 A104429
Table of sequences counting the ways of partitioning a set into length-3 arithmetic progressions, subject to various restrictions.

Generalizing further

There are lots of other generalizations that might be interesting to explore. Here’s a quick list:

  • Look at partitions of \(\{1,2,\dots,kn\}\) into \(n\) parts, all of which are an arithmetic sequence of length \(k\).
  • Count partitions of \(\{1,2,\dots,n\}\) into any number of parts of (un)equal size in a way that is (non-)self-conjugate and/or (in)separable.
  • Consider at partitions of \(\{1,2,\dots,3n\}\) into \(n\) parts, all of which are an arithmetic sequence of length \(3\), and whose diagram is “non-crossing”, that is, none of the line segments overlap anywhere. (See the 6th and 11th cases in the example for \(A279197(6) = 11\).)

If explore any generalizations of this problem on your own, if you’d like to explore together, or if you have an anecdotes about Richard Guy that you’d like to share, let me know on Twitter!

Polytopes with Lattice Coordinates

Problems 21, 66, and 116 in my Open Problem Collection concern polytopes with lattice coordinates—that is, polygons, polyhedra, or higher-dimensional analogs with vertices the square or triangular grids. (In higher dimensions, I’m most interested in the \(n\)-dimensional integer lattice and the \(n\)-simplex honeycomb).

The \(A334581(4) = 84\) ways to place an equilateral triangle on the tetrahedral grid with four points per side.
Illustration showing three of the \(\binom{7+2}{4} = 126\) equilateral triangles on a grid with seven points per side.

This was largely inspired by one of my favorite mathematical facts: given a triangular grid with \(n\) points per side, you can find exactly \(\binom{n+2}{4}\) equilateral triangles with vertices on the grid. However, it turns out that there isn’t a similarly nice polynomial description of tetrahedra in a tetrahedron or of triangles in a tetrahedron. (Thanks to Anders Kaseorg for his Rust program that computed the number of triangles in all tetrahedra with 1000 or fewer points per side.)

The \(4\)-simplex (the \(4\)-dimensional analog of a triangle or tetrahedron) with \(n-1\) points per side, has a total of \(\binom{n+2}{4}\) points, so there is some correspondence between points in some \(4\)-dimensional polytope, and triangles in the triangular grid. This extends to other analogs of this problem: the number of squares in the square grid is the same as the number of points in a \(4\)-dimensional pyramid.

The \(\binom{n+2}{4}\) equilateral triangles

I put a Javascript applet on my webpage that illustrates a bijection between size-\(4\) subsets of \(n+2\) objects and triangles in the \(n\)-points-per-side grid. You can choose different subsets and see the resulting triangles. (The applet does not work on mobile.)

The solid blue triangle corresponding to the subset \(\{4,5,8,15\} \subseteq \{1,2,\dots,16\}\).
The two smaller numbers in the subset give the size and orientation of the blue triangle, and the two larger numbers give the position.

Polygons with vertices in \(\mathbb{Z}^n\)

This was also inspired by Mathologer video “What does this prove? Some of the most gorgeous visual ‘shrink’ proofs ever invented”, where Burkard Polster visually illustrates that the only regular polygons with vertices in \(\mathbb{Z}^n\) (and thus the \(n\)-simplex honeycomb) are equilateral triangles, squares, and regular hexagons.

Polyhedra with vertices in \(\mathbb{Z}^3\)

There are some surprising examples of polyhedra in the grid, including cubes with no faces parallel to the \(xy\)-, \(xz\)-, or \(yz\)-planes.

An example of a cube from Ionascu and Obando: the convex hull of \(\{(0,3,2),(1,1,4),(2,2,0),(2,5,3),(3,0,2),(3,3,5),(4,4,1),(5,2,3)\}\)

While there are lots of polytopes that can be written with vertices in \(\mathbb{Z}^3\), Alaska resident and friend RavenclawPrefect cleverly uses Legendre’s three-square theorem to prove that there’s no way to write the uniform triangular prism this way! However, he provides a cute embedding in \(\mathbb{Z}^5\): the convex hull of $$\scriptsize{\{(0,0,1,0,0),(0,1,0,0,0),(1,0,0,0,0),(0,0,1,1,1),(0,1,0,1,1),(1,0,0,1,1)}\}.$$

Polygons on a “centered \(n\)-gon”

I asked a question on Math Stack Exchange, “When is it possible to find a regular \(k\)-gon in a centered \(n\)-gon“—where “centered \(n\)-gon” refers to the diagram that you get when illustrating central polygonal numbers. These diagrams are one of many possible generalizations of the triangular, square, and centered hexagonal grids. (Although it’s worth noting that the centered triangular grid is different from the ordinary triangular grid.)

If you have any ideas about this, let me know on Twitter or post an answer to the Stack Exchange question above.

A catalog of polytopes and grids

On my OEIS wiki page, I’ve created some tables that show different kinds of polytopes in different kinds of grids. There are quite a number of combinations of polygons/polyhedra and grids that either don’t have an OEIS sequence or that I have been unable to find.

  Square Rectangular Centered Square Triangular Centered Hexagonal
Equilateral Triangle A000332 A008893
Square A002415 A130684 A006324
Regular Hexagon A011779 A000537
Regular Polygon A002415 A130684 A006324  ? A339483*
Triangle A045996 A334705  ?  ? A241223
Rectangle A085582 A289832  ?
Right Triangle A077435  ?  ?  ? A241225
OEIS sequences for polygons on 2-dimensional grids.
Sequences marked with “*” are ones that I’ve authored, cells marked with “—” have no polygons, and cells marked with “?” do not have a corresponding sequence that I know of.
CubicTetrahedralOctahedral
Equilateral TriangleA102698A334581* A342353*
SquareA334881*A334891* ?
Regular HexagonA338322* ? ?
Regular PolygonA338323* ? ?
Triangle ? ? ?
Rectangle ? ? ?
Right Triangle ? ? ?
Regular TetrahedronA103158A269747 ?
CubeA098928 ? ?
OctahedronA178797 ? ?
Platonic SolidA338791 ? ?
OEIS sequences for polytopes on 3-dimensional grids.
Sequences marked with “*” are ones that I’ve authored, and cells marked with “?” do not have a corresponding sequence that I know of.

If you’re interested in working on filling in some of the gaps in this table, I’d love it if you let me now! And if you’d like to collaborate or could use help getting started, send me a message on Twitter!

Stacking LEGO Bricks

Back in May, I participated in The Big Lock-Down Math-Off from The Aperiodical. In the Math-Off, I went head-to-head against Colin Beveridge (who has, hands-down, my favorite Twitter handle: @icecolbeveridge). Colin wrote about using generating functions to do combinatorics about Peter Rowlett’s toy Robot Caterpillar. Coincidentally and delightfully, I wrote about using generating functions to do combinatorics about Peter Kagey’s toy LEGOs.

Counting LEGO configurations is a problem dating back to at least 1974, when Jørgen Kirk Kristiansen counted that there are 102,981,500 ways to stack six 2×4 LEGOs of the same color into a tower of height six. According to Søren Eilers, Jørgen undercounted by 4!

Animated GIF of rotating LEGO stack
One of the 102,981,504 ways to stack six 2×4 LEGOs into a tower of height six.

In my Math-Off piece, I wrote about a fact that I learned from Math Stack Exchange user N. Shales—a fact that may be my (and Doron Zeilberger’s) favorite in all of mathematics: there are exactly \(3^{n-1}\) ways to make a tower out of \(1 \times 2\) LEGO bricks following some simple and natural rules. Despite this simple formula, the simplest known proof is relatively complicated and uses some graduate-level combinatorial machinery.

The Rules

  1. The bricks must lie in a single plane.
  2. No brick can be directly on top of any other.
  3. The bottom layer must be a continuous strip.
  4. Every brick that’s not on the bottom layer must have at least one brick below it.
An tower that violates rule 1.
A tower that violates rule 2.
A tower that violates rule 3.
A tower that violates rule 4.

Gouyou-Beauchamps and Viennot first proved this result in their 1988 statistical mechanics paper, but the nicest proof that I know of can be found in Miklós Bóna’s Handbook of Enumerative Combinatorics (page 26). Bóna’s proof decomposes the stacks of blocks in a clever way and then uses a bit of generating function magic.

Other rules

In preparation for the Math-Off piece, I asked a question on Math Stack Exchange about counting the number of towers without Rule 2. The user joriki provided a small and delightful modification of Bóna’s proof that proves that there are \(4^{n-1}\) towers if only rules 1, 3, and 4 are followed.

It might also be interesting to consider the 14 other subsets of the rules. I encourage you to compute the number of corresponding towers and add any new sequences to the On-Line Encyclopedia of Integer Sequences. If you do so, please let me know! And I’d be happy to work with you if you’d like to contribute to the OEIS but don’t know how to get started.

Another natural question to ask: How many different towers can you build out of \(n\) bricks if you consider mirror images to be the same? In the example above with the red bricks, there are six different towers, because there are three pairs of mirror images. By Burnside’s Lemma (or a simpler combinatorial argument) this is equivalent to counting the number of symmetric brick stacks. If there are \(s(n)\) symmetric towers with \(n\) bricks, then there are \(\displaystyle \frac 12 (3^{n-1}+s(n))\) towers. For \(n = 4\), there are three such towers, shown below in blue.

I asked about this function on Math Stack Exchange and wrote a naive Haskell program to compute the number of symmetric towers consisting of \(n \leq 19\) bricks, which I added to the OEIS as sequence A320314. OEIS contributor Andrew Howroyd impressively extended the sequence by 21 more terms. I also added sequence \(A264746 = \frac 12 (3^{n-1}+A320314(n))\), which counts towers up to reflection, and A333650, which is a table that gives the number of towers with \(n\) bricks and height \(k\).

Stacking Ordinary Bricks

It is also interesting to count the number of (stable) towers that can be made out of ordinary bricks without any sort of mortar. I asked on Math Stack Exchange for a combinatorial rule for determining when a stack of ordinary bricks is stable. MSE user Jens commented that this problem is hard, and pointed to the OEIS sequence A168368 and the paper “Maximum Overhang” by Mike Paterson, Yuval Peres, Mikkel Thorup, Peter Winkler, and Uri Zwick, which provides a surprising example of a tower that one might expect to be stable, but in fact is not.

A surprising example of an unstable tower.
See Figure 5 of the Paterson, Peres, Thorup, Winkler, and Zwick paper.

I’d still like to find a combinatorial rule, or implement a small physics engine, to determine when a stack of bricks is stable.

These problems and some generalizations can be found in Problem 33 of my Open Problem Collection. If you’d like to collaborate on any of these problems, let me know on Twitter. If you find yourself working on your own, I’d love for you to keep me updated with your progress!

(The graphics of LEGO bricks were rendered using the impressive and free LEGO Studio from BrickLink.)