in boundendess debugging tool allow only upper or lower bounding

[mayork]

[bqpd]

What's the idea behind this?

[bqpd]

@mayork, is this issue still open? what's your desired use case for this?

[mayork]

sometimes when models are dual infeasible they will solve in a BCS but none of the variables will be near the upper or lower bound. At one point I encountered this and lower bounded all the variables. It ended up being informative.

I haven't really looked at the BCS code but I thought maybe a keyword would allow for only upper or lower obunding

[bqpd]

yeah! this will be an easy thing to add.

[whoburg]

I have a vague recollection that we found and fixed an early bug in Bounded in which we weren't properly looping through the nu's and/or lambda's to generate the final "sensitive to" printout. I wonder if maybe that could explain the (odd) behavior @mayork observed where using only an upper or lower bound was somehow more informative.

[bqpd]

This exists in #1021! Syntax is Bounded(cset, lower=0.01), etc