A Knowledge Compilation Take on Binary Polynomial Optimization

24 mai 24

Speaker: Silvia Di Gregorio

Room: B508

Date: 24/05/2024 14:00

Abstract: The Binary Polynomial Optimization (BPO) problem is defined as the problem of maximizing a given polynomial function over all binary points. The main contribution of this paper is to draw a novel connection between BPO and the problem of finding the maximal assignment for a Boolean function with weights on variables. This connection allows us to give a strongly polynomial algorithm that solves BPO with a hypergraph that is either β-acyclic or with bounded incidence treewidth. This result unifies and significantly extends the known tractable classes of BPO. The generality of our technique allows us to deal also with extensions of BPO, where we enforce extended cardinality constraints on the set of binary points, and where we seek k best feasible solutions. We also extend our results to the significantly more general problem where variables are replaced by literals. Preliminary computational results show that the resulting algorithms can be significantly faster than current state-of-the-art.

Joint work with Florent Capelli and Alberto Del Pia