gang of 44 , gang of 416

Everything about Sudoku that doesn't fit in one of the other sections

Re: gang of 44 , gang of 416

Postby gsf » Wed Jun 09, 2010 4:14 pm

ronk wrote:gsf, shouldn't -gtm.0 -e"uniq(%#mc)" on the template below generate all 44 subgrids :?: I get only one (using version 20090116).

yes it should
but there is something going on because my latest version (not posted yet) gives a different result from yours but still wrong
its on the todo list ...
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Re: gang of 44 , gang of 416

Postby gsf » Thu Jun 10, 2010 5:18 am

gsf wrote:
ronk wrote:gsf, shouldn't -gtm.0 -e"uniq(%#mc)" on the template below generate all 44 subgrids :?: I get only one (using version 20090116).

yes it should
but there is something going on because my latest version (not posted yet) gives a different result from yours but still wrong
its on the todo list ...

found the problem
the subgrid canonicalization algorithm (thanks to Michael Deverin on the programmers' forum a few years back)
has a hard coded internal stack data limit -- I upped that and it works
solver 2010-06-09 posted
there are 2 ways to get the 416 gangsters: list the solver table or make the solver generate it
Code: Select all
sudoku -gB
sudoku -gtm.0 -e'uniq(%.27#mc)' -f- -Ff%#uD m.dat

this lists the gang of 44 labeled with the lowest 416 gangster member for each 44 class
Code: Select all
sudoku -gB44

thanks for pointing the generation solution
the %.27#mc is a new hook to an older internal algorithm
its a more efficient band canonicalization algorithm and also only lists the first band (27 cells) of the grid
%.54#mc canonicalizes the first 2 bands
-Ff%#uD dumps the uniq() splay tree in order after all processing is done
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Re: gang of 44 , gang of 416

Postby dukuso » Thu Jun 10, 2010 5:53 am

how to order the gangsters ?
for most tasks it hould be best to have the 416 grouped by their 44-class

and the 44 maybe by some properties

so we must distinguish and name the orderings , and a program to easily convert
dukuso
 
Posts: 479
Joined: 25 June 2005

Re: gang of 44 , gang of 416

Postby gsf » Thu Jun 10, 2010 6:35 am

dukuso wrote:how to order the gangsters ?
for most tasks it hould be best to have the 416 grouped by their 44-class

and the 44 maybe by some properties

so we must distinguish and name the orderings , and a program to easily convert

I ordered the 416 row-order-minlex ascending and labeled them 001..416
the label for each gang of 44 element is the label of the highest labeled 416 member (don't remember why it wasn't the lowest)

we could label the gang of 44 01..44 by assigning 01 to the first 416 band and bumping the index for each successive band in a new 44 class
then two-digit would mean gang of 44 ordinal and three digit would mean 416 gangster ordinal

do you have functions that map band => 416 ordinal and band => 44 ordinal?
gsf
2014 Supporter
 
Posts: 7306
Joined: 21 September 2005
Location: NJ USA

Re: gang of 44 , gang of 416

Postby dukuso » Thu Jun 10, 2010 7:28 am

my labeling is in post 200714 above, with the assignment of the 416s to their mother 44s.

maybe call them 01a,01b,02a,...43f,44a,44b

... in _one_ of the labeling systems, to be included in the conversion program


classify the gangsters uniquely by their properties
dukuso
 
Posts: 479
Joined: 25 June 2005

Re: gang of 44 , gang of 416

Postby dukuso » Thu Jun 10, 2010 7:51 am

let mcp be an operation on a sudokugrid or band or sudoku or sub-... which consists of
permuting the 3 entries in the 9 minicolumns of a band, for all 9 minicolumns independently,
so to obtain another valid sudokugrid,band,sudoku,...
This operation may change the S-class, but keeps the 3-tuple of 44-gangsters of the bands ("G-class")


for each of the 10895 in sudokugrids occurring ordered triples of 44-band-indices ("3x44"-clases ?) count
the number of G-classes and by mcp generated different T-classes and S-classes
dukuso
 
Posts: 479
Joined: 25 June 2005

Previous

Return to General