|
| |
|
|
A090318
|
|
a(n) = least positive k such that k, k+1, k+2, ... k+n-1 is a stapled interval of length n, or 0 if no such sequence exists.
|
|
6
| |
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2184, 27829, 27828, 87890, 87890, 171054, 171054, 323510, 127374, 323510, 151062, 151062, 151062, 151061, 151060, 151059, 151058, 7106718, 7106718, 7567747, 7567746, 7567745, 7567744, 7567743, 7567742, 48595315, 48595314, 48595313
(list; graph; refs; listen; history; internal format)
|
|
|
|
OFFSET
| 1,17
|
|
|
COMMENTS
| A finite sequence of n consecutive positive integers is called "stapled" if each element in the sequence is not relatively prime to at least one other element in the sequence. Thus a staple joins two terms of the sequence whose gcd is > 1.
It has been proved that stapled intervals of length n >= 17 exist for all n.
Comments from Max Alekseyev (maxale(AT)gmail.com), Jul 24 2007: (Start)
An interval is stapled if for every element x there is another element y (different from x) such that gcd(x,y)>1.
The shortest stapled interval has length 17 and starts with the number 2184.
It is interesting to notice that the intervals [27829,27846] and [27828,27846] are stapled while the interval [27828,27845] is not.
It is clear that a stapled interval [a,b] may not contain a prime number greater than b/2 (as such a prime would be coprime to every other element of the interval).
Together with Bertrand's Postulate this implies a>b/2 or b<2a. And it follows that
* a stapled interval may not contain prime numbers at all;
* for any particular positive integer a, we can determine if it is a starting point of some stapled interval. (End)
For n>=17, a(n) < A034386(n-1) = (n-1)#. [From Max Alekseyev, Oct 08 2007]
|
|
|
REFERENCES
| A. Brauer, On a Property of k Consecutive Integers, Bull. Amer. Math. Society, vol. 47, 1941, pp. 328-331.
R. B. Eggleton, Discrete Math. 65 (1987), 141-147.
R. J. Evans, On Blocks of N Consecutive Integers, Amer. Math. Monthly, vol. 76, 1969, pp. 48-49.
R. J. Evans, On N Consecutive Integers in an Arithmetic Progression. Acta Sci. Math. Univ. Szeged, vol. 33, 1972, pp. 295-296.
I. Gassko, Common factor graphs of stapled sequences, Proceedings of the Twenty-eighth Southeastern International Conference on Combinatorics, Graph Theory and Computing (Boca Raton, FL, 1997). Congr. Numer. 126 (1997), 163-173.
I. Gassko, Stapling and composite coverings of natural numbers, Proceedings of the Twenty-seventh Southeastern International Conference on Combinatorics, Graph Theory and Computing (Baton Rouge, LA, 1996). Congr. Numer. 118 (1996), 109-116.
Heiko Harborth, Eine Eigenschaft Aufeinanderfolgender Zahlen. Arch. Math. (Basel), vol. 21, 1970, pp. 50-51.
H. L. Nelson, There is a better sequence, Journal of Recreational Mathematics, Vol. 8(1), 1975, pp. 39-43.
S. S. Pillai, On m Consecutive Integers I, Proc. Indian Acad. Sci., Sect. A, vol. 11, 1940, pp. 6-12; II ibid., vol. 11, 1940, pp. 73-80, III ibid, vol. 13, 1941, pp. 530-533; IV Bull. Calcutta Math. Soc. 36, 1944, pp. 99-101.
|
|
|
LINKS
| Max Alekseyev and William Rex Marshall, Table of n, a(n) for n = 1..103
I. Gassko, Stapled Sequences and Stapling Coverings of Natural Numbers, Electronic Journal of Combinatorics, Vol. 3, 1996, Paper R33.
|
|
|
EXAMPLE
| The shortest possible stapled sequence is [2184, 2185, 2186, 2187, 2188, 2189, 2190, 2191, 2192, 2193, 2194, 2195, 2196, 2197, 2198, 2199, 2200]
|
|
|
PROG
| (C++ program from Max Alekseyev, Oct 08 2007) For the current parameters it needs ~ 4GB of RAM to run smoothly.
It first precomputes the smallest prime divisor < 59 of each number below 10^9 and stores them in the array sp. Then uses these divisors to grow up from each particular n < 10^9 a stapled interval of the length at most 59. The records found are stored in the array L which is printed out at the end. It uses the following observations:
If a stapled interval contains a number t, then it also contains t-sp(t) or t+sp(t) or both.
If a stapled interval starts with n, n+1, ..., n+k then it must also contain the number m = max{ n + d + sp(n+d) : d=0..k, sp(n+d)>d }
Moreover, if m <= n+k, then the interval [n, n+k] is stapled.
#include <iostream>
#include <vector>
using namespace std;
#include <stdint.h>
#define D 59
#define N 1000000000ul
const uint32_t prime[16] = { 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53 };
int main() {
vector<uint32_t> sp(N+D);
vector<uint32_t> L(D);
for(int i=0; i<16; ++i) {
uint32_t p = prime[i];
for(uint32_t n=0; n<N+D; n+=p) {
if( sp[n]==0 ) sp[n] = p;
}
}
clog << "Init done\n";
for(uint32_t n=1; n<=N; ++n) {
uint32_t m = 1;
for(int d=0; d<D; ++d) {
uint32_t s = sp[n+d];
if(s==0) break;
if(s>d) m = max(m, d+s);
if(d>=m && L[d]==0) L[d]=n;
}
}
for(int i=1; i<D; ++i) {
cout << i+1 << "\t" << L[i] << endl;
}
return 0;
}
|
|
|
CROSSREFS
| Cf. A130170, A130171, A130173.
Sequence in context: A206057 A130173 A130170 * A130171 A194585 A107657
Adjacent sequences: A090315 A090316 A090317 * A090319 A090320 A090321
|
|
|
KEYWORD
| nonn,nice
|
|
|
AUTHOR
| William Rex Marshall (w.r.marshall(AT)actrix.co.nz), Jan 25 2004
|
|
|
EXTENSIONS
| Edited by N. J. A. Sloane (njas(AT)research.att.com), Aug 04 2007, Oct 08 2007
|
| |
|
|
