Terms and Terms Differences -> 1,2,3,4,...
Leroy Quet
qq-quet at mindspring.com
Sun May 2 21:31:12 CEST 2004
Here are some sequences (not in the EIS) which are interesting (to me)
because they involve recursive sequences and their compliments (sequences
of positive integers not in original sequence).
First, the least simply expressed, perhaps, sequence:
Let a(1) = 1;
Let a(m+1) =
a(m) +(number of elements of {a(1),a(2),..a(m)} which are <= m)
(figured by hand)
1, 2, 4, 6, 9, 12, 16, 20, 29, 34, 39,..
So, every positive integer is either of the form
a(m) +m-1
or
a(m+1)-a(m)+m,
but not of both forms.
a(m) +m-1: 1, 3, 6, 9, 13, 17, 22, 27, 32,..
a(m+1) -a(m) +m: 2, 4, 5, 7, 8, 10, 11, 12, 14,..
-
b(1) = 1;
for m >= 2,
(b(m+1)-b(m)) is lowest positive integer not among
{b(1),b(2),..,b(m)} or among {b(2)-b(1),b(3)-b(2),...,b(m)-b(m-1)}.
b(m): 1, 3, 7, 12, 18, 26, 35, 45,..
b(m+1)-b(m): 2, 4, 5, 6, 8, 9, 10,..
This sequence IS in the EIS.
http://www.research.att.com/projects/OEIS?Anum=A005228
But if we instead take the absolute value of the difference, and require
the c-sequence to descend whenever possible, each term remaining unique
and positive, we get:
c(1) = 1;
for m >= 2,
|c(m+1)-c(m)| is lowest positive integer not among
{c(1),c(2),..,c(m)}
or among {|c(2)-c(1)|,|c(3)-b(2)|,..,|c(m)-c(m-1)|}.
c(m): 1, 3, 7, 12, 18, 10, 19, 30, 17,..
|c(m+1)-c(m)|: 2, 4, 5, 6*, 8, 9, 11, 13,..
*(note that subtracting 6 from 12 would have gotten another 6, which is
forbidden.)
But I wonder if this c-sequence is infinite. It might, since we must
subtract whenever we can, have 2 subtractions in a row, and then the
sequence could get "caught", where neither adding nor subtracting the
lowest unused integer gets an unused integer.
We could, in this case, just require that there be at most one
subtraction in a row.
(But I do not know, with the few terms I have calculated by hand, if
there is ever {with the original rules} more than one subtraction in a
row anyway.)
thanks,
Leroy Quet
More information about the SeqFan
mailing list