Observation (a) is a direct result of the properties of dihedral groups. All 
permutations can be generated by a single rotation and a single reflection. 
https://groupprops.subwiki.org/wiki/Dihedral_group

If instead of i. 4, the original question contained i. 5, with similar 
adjacency constraints, there would be 10 permutations, generated by a rotation 
of a  pentagon, and a reflection of a pentagon, not necessarily unique. 

My solution generates the required permutations for any polygon, which is a 
generalization of the original question.



⁣Sent from BlueMail ​

On May 26, 2020, 11:28 PM, at 11:28 PM, 'Mike Day' via Programming 
<[email protected]> wrote:
>Yes, I tend to agree.  
>Brian’s enquiry limits the choice to 8 out of the 24 perms, so filter
>on all is eminently simplest.
>
>It was, however, quite interesting to (a) find that (any) one solution
>can spawn the other 7 by rotations and reverses, and (b) find a
>constructive method, albeit rather naive, for finding one solution.
>
>I wondered earlier, but didn’t ask until now, why J was necessary for
>such a small problem.  Has Brian got a monster version of this teaser
>lurking in the murky shadows?
>
>Cheers,
>
>Mike
>
>
>Sent from my iPad
>
>> On 26 May 2020, at 15:08, David Lambert <[email protected]> wrote:
>> 
>> For small sets I'd start with the permutations, then select those I
>like.
>> This shows two anti-selectors, sum of pairs equal 3, and in r2 the
>bad guys
>> are directly spelled.
>> 
>> 
>>   permutations=: A.&i.~ !
>>   Filter=: "1(#~`)(`:6)
>>   restrictions=: 1 (-.@e.) 3 = 2&(+/\)  NB. 1 where none of the
>length 2
>> infix sums equal 3
>> 
>>   restrictions Filter permutations 4
>> 0 1 3 2
>> 0 2 3 1
>> 1 0 2 3
>> 1 3 2 0
>> 2 0 1 3
>> 2 3 1 0
>> 3 1 0 2
>> 3 2 0 1
>> 
>>   NB. directly enter those I don't want
>>   r2=: 1 -.@e. [: , (>0 3;1 2;2 1;3 0)&(E."1)
>>   (r2 Filter -: restrictions Filter)@:permutations 4
>> 1
>> 
>>> Am 25.05.20 um 13:47 schrieb Brian Schott:
>>> I want all permutations of  i. 4 for which 0 and 3 cannot be
>adjacent,
>> nor can 1 and 2.
>>
>----------------------------------------------------------------------
>> For information about J forums see
>http://www.jsoftware.com/forums.htm
>----------------------------------------------------------------------
>For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to