Negative gaps and inaccurate optimal values when the solutions are suboptimal

jump-dev / ECOS.jl

A Julia interface to the ECOS conic optimization solver

https://github.com/embotech/ecos

Other

41 stars 17 forks source link

Negative gaps and inaccurate optimal values when the solutions are suboptimal #54

Closed SamZhou1996 closed 3 years ago

SamZhou1996 commented 7 years ago

While using ECOS to solve a convex programming problem, I noticed that when the solution status is :Suboptimal, it returns a primal-dual pair that has a negative duality gap. Attached is an example of such a situation, where solve!() and conic_problem!() are called from Convex.jl.

screen shot 2017-09-08 at 18 25 21

odow commented 3 years ago

Closing as stale. We now have better status reporting mechanisms.