Find which constrains are violated in an IPOPT JuMP model

I have a model I’m trying to optimise which is returning “LOCALLY_INFEASIBLE”. When using a linear solver like Gurobi at this point I’d compute conflicts, I appreciate this isn’t possible when working with a non-linear Ipopt model. But is it possible to easily find which constraints were being violated at the last performed iterations? E.g.

constraint_violation(model)

See

https://jump.dev/JuMP.jl/stable/manual/solutions/#Checking-feasibility-of-solutions

1 Like

Thanks, it turns out I was on an older version of JuMP which didn’t have this function.

Note to future me: Ensure primal_feasibility_report is using a similar atol to your solver…

1 Like