search

vgupta1 / DDUS.jl

Data-driven Uncertainty Sets for the JuMPeR framework
Other
11 stars 3 forks source link

When using UCSOracle, with no support, and a cutsolver set, cutsolver is not used #8

Closed IainNZ closed 9 years ago

IainNZ commented 9 years ago

I'm guessing because of something happening inside kappa(eps_k)?

vgupta1 commented 9 years ago

So... this is sort of by design... With no support, you can do things closed-form. It's only with support that you actually need to directly solve an optimization. Do you see a problem with this or the interface design?

On Thu Feb 19 2015 at 2:57:53 PM Iain Dunning notifications@github.com wrote:

— Reply to this email directly or view it on GitHub https://github.com/vgupta1/DDUS.jl/issues/8.

IainNZ commented 9 years ago

Ah hold on, I'm confused. I was seeing a lot of output, but I forgot I'm solving an LP, so what I was seeing is the master solving and hotstarting

IainNZ commented 9 years ago

Used to MIPs where there isn't much output if the cutsolver is set

vgupta1 commented 9 years ago

Now I'm confused. :) Do you think there is a problem or not?

On Thu Feb 19 2015 at 5:06:42 PM Iain Dunning notifications@github.com wrote:

Ah hold on, I'm confused. I was seeing a lot of output, but I forgot I'm solving an LP, so what I was seeing is the master solving and hotstarting

— Reply to this email directly or view it on GitHub https://github.com/vgupta1/DDUS.jl/issues/8#issuecomment-75171899.

IainNZ commented 9 years ago

No problem whatsover, I just confused myself