Sunday, March 21, 2010

Painting by Numbers (2)

As a follow up on Painting by numbers here is a big one:





There is also one I cannot solve quickly (may be an error in reproducing the input on my side):




To make sure we are hitting the limits of MIP solver technology here and this is not due to some transcription error in the input data I should try this data with some other solver, e.g. a constraint programming system. (I did a simple check on the input: the count of the cells covered by horizontal clusters must be the same as the count of the cells covered by the vertical clusters).