Wednesday 20 December 2017

Aliquot Sequences

My attention was drawn to aliquot sequences today, day 25098, because the number is a member of four aliquot sequences (as shown below):


From any starting point, it's easy enough to calculate the next member in the sequence by using the divisor function \(\sigma_1 \). For example, the second term in the sequence starting with 138 is \(\sigma_1 (138) -138=150\). Many sequences lead to a prime number and then terminate because \(\sigma_1 (\text{prime number}) -\text{prime number}=1\) and \(\sigma_1(1) -1=0\). The sequence beginning with 138 (OEIS A008888) has 178 members and ends in 59, 1, 0. OEIS A008889 is really the same as OEIS A008888 except for the starting point (150 instead of 138)

Aliquot sequence OEIS A008890 is different however, and starts with 168 but the second term is 312 which is the fifth term in OEIS A008888. Aliquot sequence OEIS A074907 starts with 570 but after a few terms reaches 19434 which again is a term in the OEIS A08888 sequence.

Not all aliquot sequences end. To quote from Wikipedia:
There are a variety of ways in which an aliquot sequence might not terminate:
  • A perfect number has a repeating aliquot sequence of period 1. The aliquot sequence of 6, for example, is 6, 6, 6, 6, ... 
  • An amicable number has a repeating aliquot sequence of period 2. For instance, the aliquot sequence of 220 is 220, 284, 220, 284, ... 
  • A sociable number has a repeating aliquot sequence of period 3 or greater. (Sometimes the term sociable number is used to encompass amicable numbers as well.) For instance, the aliquot sequence of 1264460 is 1264460, 1547860, 1727636, 1305184, 1264460, ... 
  • Some numbers have an aliquot sequence which is eventually periodic, but the number itself is not perfect, amicable, or sociable. For instance, the aliquot sequence of 95 is 95, 25, 6, 6, 6, 6, ... . Numbers like 95 that are not perfect, but have an eventually repeating aliquot sequence of period 1 are called aspiring numbers (OEIS  A063769).
Numbers whose Aliquot sequence is not known to be finite or eventually periodic are:
276, 306, 396, 552, 564, 660, 696, 780, 828, 888, 966, 996, 1074, 1086, 1098, 1104, 1134, 1218, 1302, 1314, 1320, 1338, 1350, 1356, 1392, 1398, 1410, 1464, 1476, 1488, ... (sequence A131884 in the OEIS) 

ADDENDUM: 17th July 2020

Today, I turned 26038 days old and I revisited OEIS A008888 (Aliquot sequence starting at 138) because this number is a member of the sequence and appears very near the end. Here is the full sequence:
138, 150, 222, 234, 312, 528, 960, 2088, 3762, 5598, 6570, 10746, 13254, 13830, 19434, 20886, 21606, 25098, 26742, 26754, 40446, 63234, 77406, 110754, 171486, 253458, 295740, 647748, 1077612, 1467588, 1956812, 2109796, 1889486, 953914, 668966, 353578, 176792, 254128, 308832, 502104, 753216, 1240176, 2422288, 2697920, 3727264, 3655076, 2760844, 2100740, 2310856, 2455544, 3212776, 3751064, 3282196, 2723020, 3035684, 2299240, 2988440, 5297320, 8325080, 11222920, 15359480, 19199440, 28875608, 25266172, 19406148, 26552604, 40541052, 54202884, 72270540, 147793668, 228408732, 348957876, 508132204, 404465636, 303708376, 290504024, 312058216, 294959384, 290622016, 286081174, 151737434, 75868720, 108199856, 101437396, 76247552, 76099654, 42387146, 21679318, 12752594, 7278382, 3660794, 1855066, 927536, 932464, 1013592, 1546008, 2425752, 5084088, 8436192, 13709064, 20563656, 33082104, 57142536, 99483384, 245978376, 487384824, 745600776, 1118401224, 1677601896, 2538372504, 4119772776, 8030724504, 14097017496, 21148436904, 40381357656, 60572036544, 100039354704, 179931895322, 94685963278, 51399021218, 28358080762, 18046051430, 17396081338, 8698040672, 8426226964, 6319670230, 5422685354, 3217383766, 1739126474, 996366646, 636221402, 318217798, 195756362, 101900794, 54202694, 49799866, 24930374, 17971642, 11130830, 8904682, 4913018, 3126502, 1574810, 1473382, 736694, 541162, 312470, 249994, 127286, 69898, 34952, 34708, 26038, 13994, 7000, 11720, 14740, 19532, 16588, 18692, 14026, 7016, 6154, 3674, 2374, 1190, 1402, 704, 820, 944, 916, 694, 350, 394, 200, 265, 59, 1, 0.
This is not the last time I'll encounter OEIS A008888 because in a couple of years time, I'll meet 26742 and 26754 (assuming I'm still alive).

Saturday 16 December 2017

Building Brilliant Numbers

I shouldn't let the opportunity go by to make a note of today and tomorrow's numbers: 25094 and 25095. Both these numbers belong to the OEIS sequence A108770: numbers n such that \(n^2 + (n+1)^2 \) is a brilliant number. The sequence proceeds:
3, 10, 15, 20, 27, 37, 59, 92, 105, 120, 152, 155, 175, 190, 215, 219, 242, 245, 254, 255, 277, 300, 302, 307, 325, 337, 362, 365, 370, 402, 415, 614, 930, 944, 987, 1049, 1059, 1112, 1192, 1204, 1210, 1220, 1265, 1312, 1344, 1360, 1374, 1449, 1460, 1504, 1527, ...
As can be seen, such numbers are relatively common. 92 is given as an example \( 92^2 + 93^2 = 17113 = 109*157 \) as both of its factors have three digits. 25094 is also in this sequence because \( 25094^2+25095^2 = 1259467861 = 23873×52757 \) but so also is 25095 because \( 25095^2+25096^2 = 1259568241 = 29401×42841 \). These consecutive pairs are relatively rare and those listed are (254,255), (4099,4100), (11159,11160), (25094, 25095), (31754,31755) and (40189,40190) with the conjecture that there are infinitely many of such pairs.

It's a bit of a wait until the next pair (31754 and 31755) but hopefully I'll be around to greet them.

Tuesday 12 December 2017

Triangulation Conjecture Disproved

This is a story from Quanta Magazine about a Romanian-American mathematician by the name of Ciprian Manolescu (Wikipedia link) who was born on December 24th 1978. He is now approaching his 39th birthday. The article appeared in January of 2015 but I only read about today via Flipboard. He managed to disprove the triangulation conjecture, which posits that all manifolds (abstract spaces) can be triangulated (one of the most famous problems in topology). He has "the sole distinction of writing three perfect papers at the International Mathematical Olympiad: Toronto, Canada (1995); Bombay, India (1996); Mar del Plata, Argentina (1997)" and "he was elected as a member of the 2017 class of Fellows of the American Mathematical Society for contributions to Floer homology and the topology of manifolds" (quoting from the Wikipedia article).

The article mentions invariants and uses the well known Euler Formula for polyhedra stating that the number of vertices V, faces F, and edges E in a convex 3-dimensional polyhedron, satisfies V + F - E = 2.

An “invariant” is a tool mathematicians use to compare spaces, or manifolds. One famous example is the Euler characteristic, shown here. To calculate it for any two-dimensional manifold, first carve the manifold into polygons (here we use triangles). Next, add the number of faces to the number of vertices and subtract the number of edges. Every sphere will have an Euler characteristic of 2, no matter how the manifold is carved up.
To quote from the article:
He created a new invariant, which he named “beta,” and used it to create a proof by contradiction. Here’s how it works: As we have seen, the triangulation conjecture is equivalent to asking whether there exists a homology 3-sphere with certain characteristics. One characteristic is that the sphere has to have a certain property — a Rokhlin invariant of 1. Manolescu showed that when a homology 3-sphere has a Rokhlin invariant of 1, the value of beta has to be odd. At the same time, other necessary characteristics of these homology 3-spheres require beta to be even. Since beta cannot be both even and odd at the same time, these particular homology 3-spheres do not exist. Thus, the triangulation conjecture is false.
Here is a link to Ciprian Manolescu's UCLA page.

Sunday 3 December 2017

Triangles with Optimal Dynamics

In my previous post, I explored the close-to-equilateral integer triangles and today I was reading another mathematical article about triangles, so-called triangles with optimal dynamics. I thought the topic worthy of a blog post. Here is a quote from the article:
When you set a ball in motion on a billiard table, it may seem as if anything is possible, but when a table has optimal dynamics, only two things truly are. The first is complete chaos, which is to say that the ball’s path will cover the entire table as time wears on. The second is periodicity — a repeating path like a ball pinging back and forth between two sides.
In tables without optimal dynamics, a wider range of possibilities exists, which makes the full analysis of all possible paths impossible: A ball could end up bouncing chaotically in one part of the table forever, never retracing its path, but also never covering the whole table.
What mathematicians do know is that there are at least eight kinds of triangles with optimal dynamics; the first was discovered in 1989 and the last in 2013. Whether there are more is anyone’s guess. 

 Let's look at each in turn:

  • 1 : 1 : n means isosceles triangles with angles 10°, 10°, 160° or 20°, 20°, 140° etc.
  • 1 : 2 : n means triangles with angles 10°, 20°, 150° or 20°, 40°, 120° etc.
  • 3 : 4 : 5 represents a triangle with angles of 45°, 60° and 75°
  • 2 : 3 : 4 represents a triangle with angles of 40°, 60° and 80°
  • 3 : 5 : 7 represents a triangle with angles of 36°, 60° and 84°
  • 1 : 4 : 7 represents a triangle with angles of 15°, 60° and 105°
  • 2 : (n-2) : n represents a right angled triangles with angles of 2°, 88° or 90°, 4°, 86° or 6°, 84°, 90° or 10°, 80°, 90° or 12°, 78°, 90° or 18°, 72°, 90° or 20°, 70°, 90° or 30°, 60°, 90° or 36°, 54°, 90°
  • 2 : n : n represents triangles like 2°, 89°,89° or 4°, 88°, 88° or 6°, 86°, 86° etc.
The article goes on to identify two quadrilaterals with optimal dynamics (shown below):


Saturday 2 December 2017

25080: A Number Of Interest

The number of OEIS entries for numbers in the region 25000 can be quite few. For example, 25079 has three entries, 25078 has six, 25077 has five and so on. Thus it was surprising to note that 25080 has an impressive 43 entries. It helps that it's highly factorisable: 2 x 2 x 2 × 3 × 5 × 11 × 19. Let's look at one of these entries.

OEIS A102341: areas of 'close-to-equilateral' integer triangles. To quote from OEIS:
A close-to-equilateral integer triangle is defined to be a triangle with integer sides and integer area such that the largest and smallest sides differ in length by unity. The first five close-to-equilateral integer triangles have sides (5, 5, 6), (17, 17, 16), (65, 65, 66), (241, 241, 240) and (901, 901, 902).
The corresponding sequence of areas for these triangles is 12, 120, 1848, 25080, 351780. So the 'close-to-equilateral' integer triangle with sides 241, 241 and 240 has an integer area of 25080. Here is the image that I posted to my Twitter account, taken from WolframAlpha:

Here is further detail from the OEIS entry:
Heron's formula: a triangle with side lengths \( (x,y,z)  \) has area \(A = \sqrt {s*(s-x)*(s-y)*(s-z)} \) where \(s = (x+y+z)/\)2. For this sequence we assume integer side-lengths \(x = y = z \pm 1 \). Then for \(A \) to also be an integer, \( x+y+z \) must be even, so we can assume \( z = 2k \) for some positive integer \( k \). Now \( s = (x+y+z)/2 = 3k \pm 1 \) and \(A = \sqrt{(3*k \pm 1)*k*k*(k \pm 1)} = k*\sqrt{3*k^2 \pm 4*k + 1} \). To determine when this is an integer, set \( 3*k^2 \pm 4*k + 1 = d^2 \). If we multiply both sides by 3, it is easier to complete the square: \((3*k \pm 2)^2 - 1 = 3*d^2 \). Now we are looking for solutions to the Pell equation \( c^2 - 3*d^2 = 1 \) with \( c = 3*k \pm 2 \), for which there are infinitely many solutions: use the upper principal convergents of the continued fraction expansion of \( \sqrt{3} \). 
 Here are links supplied in the OEIS entry:

Friday 29 September 2017

Numbers: the Early Days

As a child, certain numbers burnt their way into my memory. One of them was the number of the house that I grew up in from age two onwards. It remained the family home until the late 1980s. The house is still there in an area that is now zoned light industrial but was once residential. All the surrounding houses have long gone but the one at 21 Mayneview Street remains. See Figure 1. As it turned out, I would leave home to join the army shortly after my 21st birthday. I didn't sign up, instead I was conscripted.

Figure 1: the ancestral home as shown
on Google Earth

This was the time of the United State's intervention in Vietnam and Australia's obsequious Liberal government, eager to demonstrate its unconditional support for the Americans, was happy to supply its young males as cannon fodder. And so 21 turned out to be a turning point in my life. Thereafter, it would always be a case of life before the army and life after it. I was discharged before my 22nd birthday, so the whole "adventure" took place while I was 21 years old. 

Figure 2: my maternal grandparents's house

There are other house numbers that I recall. One was that of my maternal grandparent's house where I spent the first two years of my life. The address was 69 Hale Street, Petrie Terrace, Brisbane. See Figure 2. Around the corner lived my first friend, George Marczyk, at 56 Sheriff Street. For an overview of these locations, follow this link.

Area codes are always important and I remember that for Milton, the suburb where the house was located, the code was 4064. However, this four digit system was only introduced in 1967.

The telephone number in the family home in those early days was FM4676 because of the rotary dialer. This translates to 364676 as can be seen from the image below that also compares the system used in Australia with that in the UK. See Figure 3.

Figure 3

Clearly, it was an easy number to remember. I don't remember any other telephone numbers from that era or any telephone numbers from any era for that matter but I remember that one. Here is a little more about the system:
Older Australian rotary dial telephones also had letters, but the combinations were often printed in the centre plate adjacent to the number. The Australian letter-to-number mapping was A=1, B=2, F=3, J=4, L=5, M=6, U=7, W=8, X=9, Y=0, so the phone number BX 3701 was in fact 29 3701. When Australia changed to all-numeric telephone numbers, a mnemonic to help people associate letters with numbers was the sentence, "All Big Fish Jump Like Mad Under Water eXcept Yabbies." However, such letter codes were not used in all countries. Wikipedia.
Later, as the population grew, an extra digit was added sometime in the late 1960s or early 1970s. I'm certain the extra digit was a 9 and I'm fairly sure that it was placed after the initial 36 so that the number became 3694676.

Of course, a combination of letters and numbers is still used today with car number plates. I can remember the number plate of the first car I owned: PAE 615. It was a red, Ford Falcon station wagon of an early 60s vintage. It looked something like the car shown in Figure 4.

Figure 4

on December 29th 2020

Novel Integration Technique

How does one integrate the integral shown below? Well, the standard approaches don't work here and so a little trickery is called for. Here's the integral in question:  \[ \int_0^\infty \! \frac{\sin x}{x} \mathrm{d}x \]Firstly, let's define a function \( I(b) \) as follows:\[ I(b)=\int_0^\infty \! \frac{\sin x}{x}\mathrm{e}^{-bx} \mathrm{d}x \; \text{where } b>=0 \]Now let's differentiate both sides with respect to \(b \), not \( x \):\[ I'(b)=-\int_0^\infty  \! \sin x \, \mathrm{e}^{-bx} \mathrm{d}x \; \]Using integration by parts, \(I(b) \) can be expressed as follows: \[ I'(b)=\left. \frac{\mathrm{e}^{-bx}(\cos x + b \sin x)}{b^2+1} \right| _{x=0} ^{x=\infty} =-\frac{1}{b^2+1} \]The demonstration of this integration by parts can be found here. Now integrating both sides with respect to b, we get:\[I(b)=-\int \!\frac{1}{b^2+1} \mathrm{d}b =-\arctan b+\mathrm{C} \]Comparing this result for \( I(b) \) with our earlier result, we can write:\[ -\arctan b +\mathrm{C}=\int_0^\infty \! \frac{\sin x}{x}\mathrm{e}^{-bx} \mathrm{d}x \]Now as \( b \rightarrow \infty \), the equation reduces to: \[ -\frac{\pi}{2} + \mathrm{C}=0 \text{ and so } \mathrm{C}=\frac{\pi}{2} \]Finally, we can write:\[I(b)=-\arctan(b)+\frac{\pi}{2}=\int_0^\infty \! \frac{\sin x}{x}\mathrm{e}^{-bx} \mathrm{d}x \]Now setting \( b \)=0, we achieve our desired result, namely:\[ I(0)=\int_0^\infty \! \frac{\sin x}{x}\mathrm{d}x=\frac{\pi}{2} \]A video demonstration of this technique (known as Feynman's Technique, presumably after the famous physicist) can be found here:

Thursday 6 July 2017

From Binomial to Poisson

There's nothing special about this post. Once again, I'm practising my LaTeX skills and reminding myself at the same time of the connection between the Binomial and Poisson probability distributions. Let's start with a binomial distribution with number of trials \( n \) and probability of success \( \lambda/n \). Let's say that \( n \) is the number of equal intervals into which a unit of time has been divided. The unit of time might be one hour and this interval is divided into 10 equal intervals each of 6 minutes duration. Suppose that there are on average 7 calls during the hour. The probability of a call being received during any one of the intervals is thus 7/10 or 0.7.

Suppose X is a random variable that is binomially distributed with probability of success \( \lambda/n \) and number of trials \( n \), that is \(X \sim B(n, \lambda/n) \). The probability that X takes a special value \( x \) is then given by:\[ P(X=x)=\binom{n}{x} \left( \frac{ \lambda}{n} \right)^x \left( 1-\frac{\lambda}{n} \right)^{n-x} \]This can be written as:\[ P(X=x)=\frac{n(n-1)(n-2) ... (n-x+1)}{x!} \times \frac{\lambda^x}{n^x}\left(1-\frac{\lambda}{n} \right)^{n-x} \]Again, this can be rearranged to:\[ P(X=x)=\frac{\lambda^x}{x!} \times \frac{n-1}{n} \times \frac{n-2}{n} \times ... \times \frac{n-x+1}{n} \times \left(1-\frac{\lambda}{n} \right)^{n-x} \]An \( n  \to \infty \), the expression becomes: \[P(X=x)=\frac{\lambda^x}{x!}e^{-\lambda} \]This formula then represents the Poisson distribution, a suitable model for events which:

  • occur randomly in space or time
  • occur singly, that is events cannot occur simultaneously
  • occur independently
  • occur at a constant rate, that is a mean number of events in a given time interval is proportional to the size of the interval

A discrete random variable that follows a Poisson distribution with parameter \( \lambda \) is written as \(X \sim Po(\lambda) \) and the mean of this distribution is \( \lambda \) and the variance is also \( \lambda \). Thus the mean and variance of a Poisson distribution are equal.

In Google Sheets, the syntax for the Poisson distribution function is shown below: