Different upper and lower bound: Difference between revisions
m →Status |
Tomtom2357 (talk | contribs) No edit summary |
||
| (3 intermediate revisions by 3 users not shown) | |||
| Line 5: | Line 5: | ||
<math>N(0, b) = b</math> (everything must be <math>+1</math>) | <math>N(0, b) = b</math> (everything must be <math>+1</math>) | ||
<math>N(1, 1) = 11</math> (there are <math>4</math> such sequences: choose <math>x_1</math>, and use the constraints <math>x_n +x_{2n} = 0</math> and <math>x_1 + \ldots + x_{2n} = 0</math> to determine the entries up to <math>10</math>; then choose <math>x_{11}</math>) | <math>N(1, 1) = 11</math> (there are <math>4</math> such sequences: choose <math>x_1</math>, and use the constraints <math>x_n +x_{2n} = 0</math> and <math>x_1 + \ldots + x_{2n} = 0</math> to determine the entries up to <math>10</math>; then choose <math>x_{11}</math>) (alternatively: Our first placement can be arbitrary, so suppose 6 is +. This forces 12-, which forces 3- and 9+, which forces 2- and 4+, which forces 1+ and 8-, which forces 5- and 10+, which forces 7+, and a contradiction when d = 1 and k = 10. The sequence is thus: + – - + – + + – + + . -) | ||
<math>N(1, 2) = 41</math> (there are <math>4</math> such sequences -- example below) | <math>N(1, 2) = 41</math> (there are <math>4</math> such sequences -- example below) | ||
| Line 15: | Line 15: | ||
<math>N(1,b)<\infty</math> ([http://www.dpmms.cam.ac.uk/~ardm/erdoschu.pdf On a conjecture of Erdős and Čudakov]) | <math>N(1,b)<\infty</math> ([http://www.dpmms.cam.ac.uk/~ardm/erdoschu.pdf On a conjecture of Erdős and Čudakov]) | ||
<math>N( | Question: is <math>N(1,b)</math> always prime? | ||
For the zero-based problem, see [http://gowers.wordpress.com/2010/01/06/erdss-discrepancy-problem-as-a-forthcoming-polymath-project/#comment-4822 this | <math>N(2, 2) = 1160</math> (e.g [[the first 1124-sequence]]) | ||
For the zero-based problem, see [http://gowers.wordpress.com/2010/01/06/erdss-discrepancy-problem-as-a-forthcoming-polymath-project/#comment-4822 this comment] | |||
==Method== | ==Method== | ||
| Line 60: | Line 62: | ||
+ + + - + + - - + - - + | + + + - + + - - + - - + | ||
<math>N(2, 2) | <math>N(2, 2) = 1160</math> | ||
--[[User:Alec|Alec]] 13:46, 9 January 2010 (UTC) | --[[User:Alec|Alec]] 13:46, 9 January 2010 (UTC) | ||
==Relevant code== | ==Relevant code== | ||
The code(s) (or a link to the code(s)) used to find this sequence should be posted here. | The code(s) (or a link to the code(s)) used to find this sequence should be posted here. | ||
Latest revision as of 22:48, 30 April 2015
If [math]\displaystyle{ N(a,b) }[/math] is the maximum length of a [math]\displaystyle{ \pm 1 }[/math] sequence with the partial sums along its HAPs bounded below by [math]\displaystyle{ -a }[/math] and above by [math]\displaystyle{ b }[/math], then:
[math]\displaystyle{ N(a, b) = N(b, a) }[/math]
[math]\displaystyle{ N(0, b) = b }[/math] (everything must be [math]\displaystyle{ +1 }[/math])
[math]\displaystyle{ N(1, 1) = 11 }[/math] (there are [math]\displaystyle{ 4 }[/math] such sequences: choose [math]\displaystyle{ x_1 }[/math], and use the constraints [math]\displaystyle{ x_n +x_{2n} = 0 }[/math] and [math]\displaystyle{ x_1 + \ldots + x_{2n} = 0 }[/math] to determine the entries up to [math]\displaystyle{ 10 }[/math]; then choose [math]\displaystyle{ x_{11} }[/math]) (alternatively: Our first placement can be arbitrary, so suppose 6 is +. This forces 12-, which forces 3- and 9+, which forces 2- and 4+, which forces 1+ and 8-, which forces 5- and 10+, which forces 7+, and a contradiction when d = 1 and k = 10. The sequence is thus: + – - + – + + – + + . -)
[math]\displaystyle{ N(1, 2) = 41 }[/math] (there are [math]\displaystyle{ 4 }[/math] such sequences -- example below)
[math]\displaystyle{ N(1, 3) = 83 }[/math] (there are [math]\displaystyle{ 216 }[/math] such sequences -- example below)
[math]\displaystyle{ N(1, 4) = 131 }[/math] (there are [math]\displaystyle{ 87144 }[/math] such sequences -- example below)
[math]\displaystyle{ N(1,b)\lt \infty }[/math] (On a conjecture of Erdős and Čudakov)
Question: is [math]\displaystyle{ N(1,b) }[/math] always prime?
[math]\displaystyle{ N(2, 2) = 1160 }[/math] (e.g the first 1124-sequence)
For the zero-based problem, see this comment
Method
Here should be a short description of the way the sequences was found. (The code(s) used should be further down this page.)
Status
Is the data still relevant (e.g. longest know)? Is the method still relevant, or have we found a better method? Is the program still running on a computer somewhere?
The data
[math]\displaystyle{ N(1, 2) = 41 }[/math] (there are [math]\displaystyle{ 4 }[/math] such sequences -- example below)
0 + - - + - + + - + + - - + - + + - + - - - + - + + - - + + - + - + + - - - + - + +
[math]\displaystyle{ N(1, 3) = 83 }[/math] (there are [math]\displaystyle{ 216 }[/math] such sequences -- example below)
0 - + - + - + + - + + - - + - + + - - + - - + - + + - + + + - - - + + - + + - - + - + - - + + - - + - + + + - + - - - + + - + - + - - + - + + - + + - - + - + - - - + +
[math]\displaystyle{ N(1, 4) = 131 }[/math] (there are [math]\displaystyle{ 87144 }[/math] such sequences -- example below)
0 + - - + - + + - - + - + + - + + + + - - - + - - + - + + - - + + + - - - + + - + - + - - + + - + + - - + - - + - + + + + - - + - + - - + + + - + - - - - + + - - - + + - - + + + + - - - - + + - + - + + + + - - + + - + - - - + + - - - + - - + + + - + + - - + - - +
[math]\displaystyle{ N(2, 2) = 1160 }[/math]
--Alec 13:46, 9 January 2010 (UTC)
Relevant code
The code(s) (or a link to the code(s)) used to find this sequence should be posted here.
