Most "compact" sequence such that there is at least one prime between a(n) and a(n+1)
Hugo Pfoertner
all at abouthugo.de
Thu Dec 13 12:51:41 CET 2007
>My apologies.
>
>The problem is to find exponent E such that interval (n^E, (n+1)^E] includes a
>prime for every n >= 1. The algorithm finds.
>
>I wrote a lazy binary search which found E = log(541)/log(52), but this value
>appears to be incorrect.
>
>I wrote a more methodical algorithm to find E, which finds
>
> E = log(1151)/log(95) = 1.5477771087+
>
>The algorithm shows that no exponent e < E is admissible. Also, if we compare
>the size of the interval (n^E, (n+1)^E] to size of prime gaps listed in
>A005250 and A002386, we find it is much larger, indicating that we will find
>primes on (n^E, (n+1)^E] up to n^E = 2*10^16. Also, the prime gaps seem to be
>growing more slowly than (n+1)^E-n^E, so we can have a very warm fuzzy that E
>is the correct value.
>
>Maybe someone could verify my findings.
The decisive spot occurs at the prime 1151, which determines the least possible exponent. A description of a potential new sequence would be
floor(n^(log(1151)/log(95))) together with a remark like "there is at least one prime p in each interval fulfilling a(n) < p <= a(n+1)"
The sequence would be
1 2 5 8 12 16 20 24 29 35 40 46 52 59 66
73 80 87 95 103 111 119 128 136 145 154 164 173 183 193
203 213 224 234 245 256 267 278 290 301 313 325 337 349 362
374 387 400 413 426 439 452 466 480 493 507 522 536 550 565
579 594 609 624 639 655 670 685 701 717 733 749 765 781 798
814 831 848 865 882 899 916 933 951 968 986 1004 1022 1040 1058
1076 1095 1113 1132 1151 1169 1188 1207 1226 1246 1265 1284 1304 1324 1343
1363 1383 1403 1423 1444 1464 1485 1505 1526 1547 1567 1588 1609 1631 1652
1673 1695 1716 1738 1760 1781 1803 1825 1848 1870 1892 1915 1937 1960 1982
2005 2028 2051 2074 2097 2120 2144 2167 2191 2214 2238 2262 2286 2309 2334
and an example for the first occurring prime p within the intervals
n a(n) p a(n+1)
1 1 2 2
2 2 3 5
3 5 7 8
4 8 11 12
5 12 13 16
6 16 17 20
7 20 23 24
8 24 29 29
9 29 31 35
10 35 37 40
11 40 41 46
12 46 47 52
13 52 53 59
14 59 61 66
15 66 67 73
16 73 79 80
17 80 83 87
18 87 89 95
19 95 97 103
20 103 107 111
...
the decisive part:
93 1113 1117 1132
94 1132 1151 1151
95 1151 1153 1169
Hugo
More information about the SeqFan
mailing list