Search papers, labs, and topics across Lattice.
This paper investigates the problem of constructing compressed datasets that preserve optimal decisions in linear programs with uncertain cost vectors. It proves that computing the minimum size of such "sufficient decision datasets" (SDDs) is NP-hard, and deciding global sufficiency is coNP-hard. To overcome this, the authors introduce "pointwise sufficiency," a relaxed criterion, and develop a polynomial-time cutting-plane algorithm for constructing pointwise-sufficient datasets, along with a cumulative algorithm that provides PAC guarantees in a data-driven setting.
Forget big data: you only need a tiny, decision-sufficient subset to guarantee near-optimal solutions in linear programs, even with uncertain costs.
We study how to construct compressed datasets that suffice to recover optimal decisions in linear programs with an unknown cost vector $c$ lying in a prior set $\mathcal{C}$. Recent work by Bennouna et al. provides an exact geometric characterization of sufficient decision datasets (SDDs) via an intrinsic decision-relevant dimension $d^\star$. However, their algorithm for constructing minimum-size SDDs requires solving mixed-integer programs. In this paper, we establish hardness results showing that computing $d^\star$ is NP-hard and deciding whether a dataset is globally sufficient is coNP-hard, thereby resolving a recent open problem posed by Bennouna et al. To address this worst-case intractability, we introduce pointwise sufficiency, a relaxation that requires sufficiency for an individual cost vector. Under nondegeneracy, we provide a polynomial-time cutting-plane algorithm for constructing pointwise-sufficient decision datasets. In a data-driven regime with i.i.d.\ costs, we further propose a cumulative algorithm that aggregates decision-relevant directions across samples, yielding a stable compression scheme of size at most $d^\star$. This leads to a distribution-free PAC guarantee: with high probability over the training sample, the pointwise sufficiency failure probability on a fresh draw is at most $\tilde{O}(d^\star/n)$, and this rate is tight up to logarithmic factors. Finally, we apply decision-sufficient representations to contextual linear optimization, obtaining compressed predictors with generalization bounds scaling as $\tilde{O}(\sqrt{d^\star/n})$ rather than $\tilde{O}(\sqrt{d/n})$, where $d$ is the ambient cost dimension.
