This package contains a set of example MIPs in LP format related to the paper:

"Constrained Discrete Black-Box Optimization using Mixed-Integer Programming",
T. Papalexopoulos, C. Tjandraatmadja, R. Anderson, J.P. Vielma, D. Belanger,
2021.
https://arxiv.org/abs/2110.09569

These are snapshots of the MIP acquisition problems solved in the method, in
increments of 50 steps (after the initial set of 50 points) up to 501 steps.
The problems here are the constrained random neural network with subset
equality constraints (5 instances with 3 different number of constraints each,
i.e. 15 instances total), the quadratic programs from MINLPLib (50 instances),
and NASBench101 (1 instance). This contains 3 replications for each problem.

They are otherwise set up as in the paper, although they were obtained
separately from the runs in the paper. All data used to generate these are
either randomly generated or public (in the case of MINLPLib and NASBench101;
see the paper for references).
