I have a non-convex Quadratic Programming over unite simplex set. I have a valid lower bound on the objective function (goal is minimization problem). If I add a constraint like $$f(x)\geq lower~bound,$$ the solver time is increased, since I add a non-convex constraint! is there any efficient way to use a lower bound in the input when we use Gurobi as a solver (in yalmip)?
Asked
Active
Viewed 206 times
3
-
See https://or.stackexchange.com/questions/419/feeding-known-lower-bounds-to-solvers – RobPratt Dec 23 '22 at 00:27
1 Answers
2
There's a cutoff parameter that may help. Another approach, though may or may not work is to try with $Z(x)$ $=$ lower_bound$+f(x)$
Sutanu Majumdar
- 3,461
- 1
- 2
- 11
-
I am not fully aware of that but a Gurobi engineer did post this: https://support.gurobi.com/hc/en-us/articles/360048430691-How-do-I-pass-an-objective-bound-to-Gurobi-. But then cutoff parameter can be queried and modified as another staff member confirmed here: https://support.gurobi.com/hc/en-us/community/posts/360065230572-How-to-set-objective-upper-bound- – Sutanu Majumdar Dec 23 '22 at 02:08