[seqfan] Re: D(n)

[David Sycamore]

Thanks Giovanni, that’s definitely an improvement. You did get the definition right (see drafts A330273 and A330284, where the definition of D(n) has now been improved by M.F. Hasler). 886 was my mistake, obviously 688 is better (no descending digits). However, for me 688 is a(4) not a(5), because the offset is 0. So we now have a(0) to a(9) inclusive, which is 2 up on yesterday. You have also shown that we can have numbers in which a digit occurs more than twice (in your a(9), 9 occurs three times). 
Did you use a code to find these terms (so we are sure they satisfy the minimal condition?).  If not, is there anyone out there who would like to write a code for this sequence? It would be interesting to see what further terms there are, and in particular, to know if it’s finite or not. 
Best
David 

>> On 9 Dec 2019, at 13:32, Giovanni Resta <giovanni.resta at iit.cnr.it> wrote:
>> 
>> On 12/8/19 11:07 PM, David Sycamore via SeqFan wrote:
>> 
>> Combining the above with your results, the updated terms are now (up to a(7)):
>> 1, 11, 112, 166, 886, 4468, 22468, 112468...
> 
> Uhm. If I got the definition right, a(1)-a(10) =
> 
> 
> 1, 11, 112, 166, 688, 4468, 22468, 112468, 124699, 1678999, ...
> 
> In particular,
> I got a(5)=688 (since 688 -> 616 -> 121 -> 22 -> 4) instead of a(5)=886.
> 
> Giovanni
> 
> --
> Seqfan Mailing list - http://list.seqfan.eu/