I'm working with a modeling approach (Inductive Process Modeling) that composes sets of "model structures" from specifications of "entities" and "processes". Some process combinations don't make sense, so we use sets of simple constraints in generating candidate model structures, eg:
ALWAYS-TOGETHER These processes must always be used together. ATMOST-ONE At most one of these processes may be used. EXACTLY-ONE Exactly one of these processes must be used. FORBIDDEN This process must not be used. NECESSARY All of these processes must be used. Because the number of combinations could be very large, we generate pseudo-random subsets (of a specified size). These are evaluated (eg, fitting parameters to match observed data) and the best results are presented to the user for inspection. Currently, these combinations are being produced by means of complex Lisp code that performs boolean satisfiability logic. I'm wondering whether this task (along with possible follow-on extensions involving feedback from the evaluation task) might be a good use case for Gecode. (If this let me replace the code with a small set of constraints, I'd be a really happy camper. :-) I'm a complete newbie at Gecode, so I'm hoping for caveats, comments, suggestions, etc. In addition, if any Gecoders out there are available for small consulting projects, I'd like to hear from them (off-list!). -r -- http://www.cfcl.com/rdm Rich Morin r...@cfcl.com http://www.cfcl.com/rdm/resume San Bruno, CA, USA +1 650-873-7841 Software system design, development, and documentation _______________________________________________ Gecode users mailing list users@gecode.org https://www.gecode.org/mailman/listinfo/gecode-users
