in reply to pathsearch/backtracking in Perl - how to solve a problem apparently requiring it? (comparison of Perl with Standard ML)

For comparison, here's a Haskell implementation:
change _ 0 = [[]] change [] _ = fail "can't make change" change (c:cs) n | c > n = change cs n | otherwise = map (c:) (change (c:cs) (n-c)) ++ change cs n
Unlike the original, this one finds all of the solutions for a change problem:
> change [5,2] 16 [[5,5,2,2,2],[2,2,2,2,2,2,2,2]]
Because Haskell computes lazily, if you ask for only the first solution, that's all the further Haskell searches. The following variation returns only the first solution, if any:
quickchange cs n = listToMaybe (change cs n) > quickchange [5,2] 16 Just [5,5,2,2,2] > quickchange [7] 16 Nothing
Further, we can use fractional coins if we want. How can we make change for 10/6 if we have only 1/2 and 1/3 coins?
> :m +Data.Ratio > quickchange [1%2, 1%3] (10%6) Just [1%2, 1%2, 1%3, 1%3]

Cheers,
Tom

  • Comment on Re: pathsearch/backtracking in Perl - how to solve a problem apparently requiring it? (comparison of Perl with Standard ML)
  • Select or Download Code