Closed egri-nagy closed 5 years ago
The expected timeline is possibly next week, most likely the week starting Sept 11. I'm sorry, I'll be on holidays in the coming days! I hope it's not too late until then.
No worries. I got very busy (accepted talk for clojure/conj), so it will be some time before I can return to this problem. best @
On Fri, Aug 25, 2017 at 5:18 AM, Mate Soos notifications@github.com wrote:
The expected timeline is possibly next week, most likely the week starting Sept 11. I'm sorry, I'll be on holidays in the coming days! I hope it's not too late until then.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/msoos/cryptominisat/issues/421#issuecomment-324745533, or mute the thread https://github.com/notifications/unsubscribe-auth/AGiektjo8zjZFTYPvCeFb7EYYmsz7ALJks5sbdqbgaJpZM4O9_tM .
I can't do this... I tried a few times but every time it's a mess because of variable elimination and solution extension. I don't know what to do... I would really need to think about how solution extension can be done with this in mind :S Breaking variable elimination is essentially not an option as it's super-important for speed.
I see. Thanks for the effort! Do you think it is generally true that finding a solution and finding all solutions cannot be optimized at the same time?
Well, there are exact counters out there, e.g. sharpSAT if you want to do that :) Or, much-much more interestingly, there is approxMC which is using an old CryptoMiniSat :)
I am closing this because exact counters have been improving over the years, for example, we just published work into IJCAI to improve exact counting with SharpSAT. Furthermore, approximate counting has significantly improved over the years. Getting "GreedyUnbound" right is very-very messy and currently I am not even sure it's possible. Hence, I'd rather not attempt doing it. Instead, I'll let exact counters such as SharpSAT and approximate counters such as ApproxMC do the heavy lifting.
I hope this is not too big of a disappointment and I hope you will enjoy using ApproxMC or our improved SharpSAT :) Let me know if you need any help! Cheers,
Mate
Following an off-site discussion, this issue is about the possibility of trying to unbind variables (make them Undef/) such that the solution is still valid, in the context of obtaining a large number of solutions.