Data.Algorithm.Munkres

The Munkres version of the Hungarian Method for weighted minimal bipartite matching. The implementation is based on Robert A. Pilgrim's notes, http://216.249.163.93/bob.pilgrim/445/munkres.html (mirror: http://www.public.iastate.edu/~ddoty/HungarianAlgorithm.html).

Documentation

hungarianMethodInt :: UArray (Int, Int) Int -> ([(Int, Int)], Int)

Needs a rectangular array of nonnegative weights, which encode the weights on the edges of a (complete) bipartitate graph. The indexing should start from (1,1). Returns a minimal matching, and the cost of it.

Unfortunately, GHC is opposing hard the polymorphicity of this function. I think the main reasons for that is that the there is no Unboxed type class, and thus the contexts IArray UArray e and MArray (STUArray s) e (ST s) do not know about each other. (And I have problems with the forall s part, too).

hungarianMethodFloat :: UArray (Int, Int) Float -> ([(Int, Int)], Float)

hungarianMethodDouble :: UArray (Int, Int) Double -> ([(Int, Int)], Double)

hungarianMethodBoxed :: (Real e, IArray a e) => a (Int, Int) e -> ([(Int, Int)], e)

The same as 'hungarianMethod<Type>', but uses boxed values (thus works with any data type which an instance of Real). The usage of one the unboxed versions is recommended where possible, for performance reasons.