Computing optimality certificates for convex mixed-integer nonlinear problems

Every optimization problem has a corresponding verification problem that checks whether a given optimal solution is in fact optimal. In the literature, there are a lot of such ways to verify optimality for a given solution, for example, the branch-and-bound tree. To simplify this task, optimality certificates were introduced for convex mixed-integer nonlinear programs, and it was shown that the sizes of the certificates are bounded in terms of the number of integer variables. We introduce an algorithm to compute the certificates and conduct computational experiments. Through the experiments, we show that the optimality certificates can be surprisingly small.