Projection-width: a unifying structural parameter for separable discrete optimization
Alberto Del Pia
TL;DR
The paper introduces projection-width, a unifying structural parameter for separable constraint systems, and develops polynomial-time algorithms for optimization, counting, top-$k$, and weighted constraint violation when this width is bounded. By translating general constraints into projection-aware representations and leveraging branch decompositions, it unifies tractable nonlinear discrete optimization with classical results in ILP, binary polynomial optimization, and SAT. The key contributions include new projection/shape machinery, efficient construction of projections, and four algorithmic frameworks with explicit runtimes tied to width and a membership-testing bound. These results illuminate a broad, practical tractability frontier for nonlinear discrete optimization and counting, with concrete implications for PS-width, incidence treewidth, and Fortet-linearized binary optimization.
Abstract
We introduce the notion of projection-width for systems of separable constraints, defined via branch decompositions of variables and constraints. We show that several fundamental discrete optimization and counting problems can be solved in polynomial time when the projection-width is polynomially bounded. These include optimization, counting, top-k, and weighted constraint violation. Our results identify a broad class of tractable nonlinear discrete optimization and counting problems. Even when restricted to the linear setting, they subsume and substantially extend some of the strongest known tractability results across multiple research areas: integer linear optimization, binary polynomial optimization, and Boolean satisfiability. Although these results originated independently within different communities and for seemingly distinct problem classes, our framework unifies and significantly generalizes them under a single structural perspective.