|
|
A156954
|
|
Integers N such that by inserting + or - or * or / or ^ between each of their digits, without any grouping parentheses, you can get N (the ambiguous a^b^c is avoided).
|
|
3
|
|
|
736, 2592, 11664, 15617, 15618, 15622, 15624, 15632, 15642, 15645, 15656, 15662, 15667, 15698, 17536, 27639, 32785, 39363, 39369, 45947, 46633, 46644, 46648, 46655, 46660, 46663, 117635, 117638, 117639, 117642, 117643, 117647, 117650
(list;
graph;
refs;
listen;
history;
text;
internal format)
|
|
|
OFFSET
|
1,1
|
|
COMMENTS
|
The single-digit numbers 0, ..., 9 are here excluded by convention although they also ("voidly") satisfy the definition and therefore logically should be terms of this sequence. This is in contrast to the Friedman numbers A036057 where the construction also allows concatenation of digits but then of course has to exclude the case where only concatenation of the digits is used, which excludes the single-digit terms. - M. F. Hasler, Jan 07 2015
|
|
LINKS
|
|
|
EXAMPLE
|
736 = 7 + 3^6.
2592 = 2^5*9^2.
11664 = 1*1*6^6/4.
15617 = 1*5^6 - 1 - 7.
For more examples, see the link to "decompositions".
|
|
PROG
|
(PARI) is(n, o=Vecsmall("*+-^/"))={v=Vecsmall(Str(n, n\10)); forstep(i=#v, 3, -2, v[i]=v[i\2+1]); n>9 && forvec(s=vector(#v\2, i, [1, #o-(v[i*2+1]==48)]), for(i=1, #s, 94==(v[2*i]=o[s[i]])&&i>1&&s[i-1]==4&&next(2)); n==eval(Strchr(v))&&return(1))}
|
|
CROSSREFS
|
|
|
KEYWORD
|
base,nonn
|
|
AUTHOR
|
|
|
EXTENSIONS
|
|
|
STATUS
|
approved
|
|
|
|