First, from the rec.puzzles FAQ that I maintain:
... A detailed explanation is available at
http://www.frontiernet.net/~mwdaly/recpuzzles/einstein.html. For the
record, Einstein didn't write the puzzle and far more than 2% of the
world's population could solve it.
As for your approach to solving it, it's certainly possible to permutate all possible combinations, then use perl version of the statements to remove solutions that can't work. However, technicially, I wouldn't call this 'solving' the riddle, since at some point you'd visit all 60^5 possible solutions (I believe that's the right number), and thus this is more like solving an NP problem by just trying every possibility, unlike using some heuristics to pick and choose.
(Now, with rules based programming, it technicially still would have to be solved by the same methods of visiting every solution given the way rules programming is set up, but one can argue that the rules-based system is more efficient since it might find the solution before visiting all solutions.)
-----------------------------------------------------
Dr. Michael K. Neylon - mneylon-pm@masemware.com
||
"You've left the lens cap of your mind on again, Pinky" - The Brain
"I can see my house from here!"
It's not what you know, but knowing how to find it if you don't know that's important
| [reply] |
Actually I think this is an isomorphism of Lewis Carroll's Zebra Puzzle.
You can read about the puzzle in Krzysztof Apt's lecture notes:
http://www.cwi.nl/~apt/constraints/notes1.ps
It starts at the bottom of page 8, which is really the 10th page if
you include the two title pages...
I solved this puzzle in two different ways in Prolog and Lisp.
In Prolog I chose to first specify all the constraints on the domains
of the five values, every constraint shrinking the domains and in the
end you're left with only one solution.
In Lisp I used streams to generate all permutations in a lazy way
and I filtered out all the possibilities that clashed with the rules
in the same way you'd sieve primes from the stream of integers.
The biggest choice I think is if you see the house number as a value
certain variables {colour, drink, pet, nationality, profession} can have
or if you see everything as a symbol and reason about things occurring
together or not.
I'm very curious to know how fast your Perl solution is.
I know it doesn't make much sense calling any numbers,
since there are too many variables, but my Prolog version took
0.004 seconds, using swi-prolog on a Sun Ultra 5 and I didn't time
the Lisp version. I'm just wondering if Perl will be in the same
order of magnitude.
I was also completely amazed that the Englishman doesn't drink tea
and that the Italian doesn't drink coffee, but I assumed that was just
to make things less obvious :)
Pasted in here is the Prolog version, as it's the most readable of the two:
% find the Zebra and the Water, zebra_water(X).
% by Willem Robert van Hage, wrvhage@science.uva.nl
right(2,1).
right(3,2).
right(4,3).
right(5,4).
nextdoor(X,Y) :- right(X,Y).
nextdoor(X,Y) :- right(Y,X).
match([],_).
match([X|Xs],Ys) :-
member(X,Ys),
match(Xs,Ys).
zebra_water(X) :-
nextdoor(Horse,Diplomat),
nextdoor(Fox,Doctor),
nextdoor(Norwegian,Blue),
Y = [house( Green, _, green, _, _, _),
house( White, _, white, _, _, _),
house(Norwegian, norwegian, _, _, _, _),
house( Blue, _, blue, _, _, _),
house( Fox, _, _, _, fox, _),
house( Doctor, _, _, doctor, _, _),
house( Horse, _, _, _, horse, _),
house( Diplomat, _, _, diplomat, _, _),
house( 3, _, _, _, _, milk),
house( 1, norwegian, _, _, _, _),
house( _, englishman, red, _, _, _),
house( _, spaniard, _, _, dog, _),
house( _, japanese, _, painter, _, _),
house( _, italian, _, _, _, tea),
house( _, _, yellow, diplomat, _, _),
house( _, _, green, _, _, coffee),
house( _, _, _, sculptor, snails, _),
house( _, _, _, violinist, _, juice),
house( _, _, _, _, zebra, _),
house( _, _, _, _, _, water) ],
X = [ house(1,_,_,_,_,_),
house(2,_,_,_,_,_),
house(3,_,_,_,_,_),
house(4,_,_,_,_,_),
house(5,_,_,_,_,_) ],
right(Green,White),
match(Y,X).
It should come out nicely aligned when pasted into an ascii editor :)
brother Willem (wrvhage)
--
wrvh@xs4all.nl | http://wrvh.xs4all.nl | [reply] |