LP approximations to mixed-integer polynomial optimization problems

We present a class of linear programming approximations for constrained optimization problems. In the case of mixed-integer polynomial optimization problems, if the intersection graph of the constraints has bounded tree-width our construction yields a class of linear size formulations that attain any desired tolerance. As a result, we obtain an approximation scheme for the "AC-OPF" problem on graphs with bounded tree-width. We also describe a more general construction for pure binary optimization problems where individual constraints are available through a membership oracle; if the intersection graph for the constraints has bounded tree-width our construction is of linear size and exact. This improves on a number of results in the literature, both from the perspective of formulation size and generality.

No reviews yet.

No comments yet.