Up, Up and Away With Ulam Seqeunces
This article in the ongoing series of ‘mathematical puzzles
you could investigate with a calculator’ (that’s why I just call it ‘Calculator
Fun and Games’) is the Ulam Sequence. It’s an interesting mathematical sequence
named after the Polish-American mathematician Stanisław Ulam. It’s a sequence
rather than a series, and it generates some strange and interesting results, which
seem almost random.
Take with two numbers (specifically, positive integers). A good place to start is with a =1 and
b=2. To find the next number in the
sequence, find the smallest integer that can be written as the sum of two
distinct earlier numbers in just one way.
Continue with the next number, and the next.
For example, let’s start the Ulam Sequence with the numbers
1 and 2. These are the first two terms
in the sequence.
The next term is 3 (since 1+2=3).
After that comes 4 (since 1+3=4).
The next terms is not 5. We can write 5
as 1+4 and as 2+3 using the terms that we’ve generated already.
The next term is 6 (since 2+4=6), and we can only write this in one way using
our terms.
We can write 7 as 4+3 and as 6+1, so we skip 7.
The next term is 8 (since 2+6=8).
So, the beginning of the sequence is: 1,2,3,4,6,8,…
(and it continues
1,2,3,4,6,8,11,13,16,18,26,28,36,38,47,48,53,57,62,69,72,77,82,87,97,99,102)
Note that 5, 7, 9, 10, 12, 14, 15, 17, 19, 20, 21 and 22 can be obtained in multiple
ways using the terms before them.
However: 23 is not in the sequence
because it cannot be obtained using the previous terms at all! 24 can
be written as both 16+8 and 18+6, while 25 is not obtainable.
The sequence grows in an irregular, almost random pattern. Let’s see what happens when we start with 1
and 3 instead of 1 and 2.
4 = 1 + 3 only
5 = 1 + 4 only
6 = 5 + 1 only
7 can be written as 4+3 and 6+1
8 = 5+3
9 is 4+5 and 6+3
10 = 6+4
11 is 6+5 and 8+3
The first 20 terms for the 1,3 Ulam Sequence are:
1,3,4,5,6,8,10, 12,17, 21, 23, 28, 32,34,39,43,48,52,59 and 63.
The Ulam Sequence is an interesting example of how simple
rules can lead to complex and intriguing mathematical structures, which makes
it ideal for calculator (or spreadsheet) exercises. For example, here’s a comparison of the Ulam
sequences for 1,2 compared with 1,3 (as I’ve calculated above) and then 1,4 and
1,5. Interestingly, the 1,5 sequence
does not race ahead of the 1,2 sequence as I had originally expected.
Term |
Ulam (1,2) |
Ulam (1,3) |
Ulam (1,4) |
Ulam (1,5) |
1 |
1 |
1 |
1 |
1 |
2 |
2 |
3 |
4 |
5 |
3 |
3 |
4 |
5 |
6 |
4 |
4 |
5 |
6 |
7 |
5 |
6 |
6 |
7 |
8 |
6 |
8 |
8 |
8 |
9 |
7 |
11 |
10 |
10 |
10 |
8 |
13 |
12 |
16 |
12 |
9 |
16 |
17 |
18 |
20 |
10 |
18 |
21 |
19 |
22 |
11 |
26 |
23 |
21 |
23 |
12 |
28 |
28 |
31 |
24 |
13 |
36 |
32 |
32 |
26 |
14 |
38 |
34 |
33 |
38 |
15 |
47 |
39 |
42 |
38 |
16 |
48 |
43 |
46 |
40 |
17 |
53 |
48 |
56 |
41 |
18 |
57 |
52 |
57 |
52 |
19 |
62 |
59 |
66 |
57 |
20 |
69 |
63 |
70 |
69 |
There’s a balance between the ability to leap to larger numbers (1,5) initially – from 1 to 5 – and the need to fill in more numbers between 5 and 10 (because there are very smaller numbers that can be made in multiple ways).
A quick comparison of the Ulam Sequences for (2,b) is even more interesting. We have to start with 2,3 since 2,1 is the same as 1,2 above, and 2,2 will only produce the even numbers (which is cute but dull). In fact, any even number paired with 2 will produce uninteresting results!
Let’s compare 2,3 and 2,5: These grow at a slower rate compared to the 1,b sequences. Interestingly, they contain far fewer even numbers than the 1,b sequences; in fact 2,5 only contains the even numbers 2 and 12 in the first 20 terms (with no indication that there are any more even numbers further along the sequence).
Term |
2,3 |
2,5 |
1 |
2 |
2 |
2 |
3 |
5 |
3 |
5 |
7 |
4 |
7 |
9 |
5 |
8 |
11 |
6 |
9 |
12 |
7 |
13 |
13 |
8 |
14 |
15 |
9 |
18 |
19 |
10 |
19 |
23 |
11 |
24 |
27 |
12 |
25 |
29 |
13 |
29 |
35 |
14 |
30 |
37 |
15 |
35 |
41 |
16 |
36 |
43 |
17 |
40 |
45 |
18 |
41 |
49 |
19 |
46 |
51 |
20 |
51 |
55 |
10, 11, 21, 31, 32, 41, 43, 51, 54, 61, 62, 65…
After huge initial leaps of +10 or +11 between consecutive terms, the growth rate of the sequence starts to slow down. There is only one term in the 20s, then two in the 30s, 40s and 50s, then three in the 60s.
Further reading:
Wolfram has lists and links for many of the 1,b and 2,b Ulam sequences
No comments:
Post a Comment