🚀 go-pugleaf

In article <1992Apr6.140627.10533@dcs.qmw.ac.uk> leonardo@dcs.qmw.ac.uk (Mike Hopkins) writes:


   The problem insists that each row and column contains exactly one
   queens: therefore the program should be:

   fourQueens(q(r(S11, S12, S13, S14),
		r(S21, S22, S23, S24),
		r(S31, S32, S33, S34),
		r(S41, S42, S43, S44))) :-
	   %% rows
	   bool:sat(card([1],[S11, S12, S13, S14])),
	   bool:sat(card([1],[S21, S22, S23, S24])),
	   bool:sat(card([1],[S31, S32, S33, S34])),
	   bool:sat(card([1],[S41, S42, S43, S44])),
	   %% columns
	:
	:

You have pruned the search space, by coding the fact that two queens
cannot stand on the same row or column. However, this is cheating. 

My original question was why the program was so slow. Are there any
uses of boolean constraints except for digital hardware?
--
Mattias Waldau                              
Computing Science Department                mattias@csd.uu.se
P.O. Box 520, S-751 20  Uppsala, Sweden     Phone: +46-18-181055