Hi, I’m working on a bilevel problem that is converted into an equivalent non-convex QCQP by applying strong duality to the lower-level problem (LP).
My code returns a (local) optimal solution with Ipopt solver (w/ ma86).
However, the same code returns infeasibility with Gurobi v9.0.2 (NonConvex=2).
This makes me wonder if the solution obtained with Ipopt is reliable.
Is there any possible reasoning for this? Any advice would be highly appreciated.
[a076750e] CPLEX v0.6.5 [336ed68f] CSV v0.6.2 [aaaa29a8] Clustering v0.14.1 [a93c6f00] DataFrames v0.21.3 [31c24e10] Distributions v0.23.4 [2e9cd046] Gurobi v0.8.1 [f67ccb44] HDF5 v0.13.2 [7073ff75] IJulia v1.21.2 [b6b21f68] Ipopt v0.6.2 [4076af6c] JuMP v0.21.3 [e5e0dc1b] Juno v0.8.2 [67920dd8] KNITRO v0.9.2  MAT v0.8.0 [6405355b] Mosek v1.1.3 [1ec41992] MosekTools v0.9.3 [8bb1440f] DelimitedFiles