Yuriy Zinchenko

Hi Yuheng.  Thanks.  The log and the run looks normal, as the iterates converge.  Note since we solve an embedded model, which targets both feasibility and optimality at the same time, often the co...

Thanks, Yuhen.  The log and the run is normal.Note the we solve the embedded model, which simultaneously targets feasibility and optimality, so as long as the iterates seem to converge, things are ...

Hi Yuheng.  This is a massive model, from what it sounds!To make sure I do not miss something important, could you please paste a snapshot of the log here (say, the one printed to the console)?

Thank you for the feedback, Saska.  One thing I could potentially add is that right now we deliver a small IIS with high probability, due to the nature of how things are implemented. That is, if yo...

ps: with respect how the crossover comes into play, if the model is degenerate, the algorithmic path alone will determine the values here. From a user point of view this may seem like the outcome d...

Hi Elena, to chirp in here, most likely your dual is what is called degenerate, that is, admits several solutions, in which case you may see some surpassing values coming up as well as should proba...

Hello again, Andrey. Thank you again for asking an interesting Q.  The problem instance you have shared is small, but is seemingly tough to determine feasibility.  Our apologies for not replying so...

Great to hear this is helpful. Stay well, Ramy, and please say "hello" to John if you run into him; he may remember Yuriy who was a PDF at McMaster when Professor Chinneck visited. Regards.

> If Guroubi computes IIS using a filtering algorithm, that means we can attempt to get multiple IISes by shuffling the constraints in the model, correct? I tried doing that but Guroubi's computIIS...

Hello Ramy, Thanks for asking! :) 1. For LP, computing all III's based on the dual, as you said, is equivalent to enumerating all the vertices of the respective dual polyhedron, which could be a ma...