Coresets for Clustering with General Assignment Constraints

Designing small-sized \emph{coresets}, which approximately preserve the costs of the solutions for large datasets, has been an important research direction for the past decade. We consider coreset construction for a variety of general constrained clustering problems. We significantly extend and generalize the results of a very recent paper (Braverman et al., FOCS'22), by demonstrating that the idea of hierarchical uniform sampling (Chen, SICOMP'09; Braverman et al., FOCS'22) can be applied to efficiently construct coresets for a very general class of constrained clustering problems with general assignment constraints, including capacity constraints on cluster centers, and assignment structure constraints for data points (modeled by a convex body $\mathcal{B})$. Our main theorem shows that a small-sized $\epsilon$-coreset exists as long as a complexity measure $\mathsf{Lip}(\mathcal{B})$ of the structure constraint, and the \emph{covering exponent} $\Lambda_\epsilon(\mathcal{X})$ for metric space $(\mathcal{X},d)$ are bounded. The complexity measure $\mathsf{Lip}(\mathcal{B})$ for convex body $\mathcal{B}$ is the Lipschitz constant of a certain transportation problem constrained in $\mathcal{B}$, called \emph{optimal assignment transportation problem}. We prove nontrivial upper bounds of $\mathsf{Lip}(\mathcal{B})$ for various polytopes, including the general matroid basis polytopes, and laminar matroid polytopes (with better bound). As an application of our general theorem, we construct the first coreset for the fault-tolerant clustering problem (with or without capacity upper/lower bound) for the above metric spaces, in which the fault-tolerance requirement is captured by a uniform matroid basis polytope.