sisl/NeuralVerification.jl

zero_objective! not necessary

Closed this issue · 3 comments

I'm pretty sure solving a problem without specifying an objective simply solves a feasibility problem, so it should be equivalent to specifying @objective(m, Max, 0), which is what we do now.

Yes. The purpose of having zero_objective! is just to make it more clear that we are solving a feasibility problem.

Ah, I see. In that case we can be even more explicit with something like feasibility_problem!(model)

Even better!