Reproduction code for a router that projects samples onto nonlinear, discrete, and output-coupled constraints over a frozen neural-network surrogate — feasible-by-construction optimization over NN surrogates, benchmarked on real-world data
neural-network reproducible-research constrained-optimization projection constraint-handling-rules opennn surrogate-optimization mixed-integer-optimization mixed-integer constraint-handling
-
Updated
Jul 20, 2026 - Python