English

A Tight Bound for Stochastic Submodular Cover

Data Structures and Algorithms 2021-08-03 v2 Artificial Intelligence

Abstract

We show that the Adaptive Greedy algorithm of Golovin and Krause (2011) achieves an approximation bound of (ln(Q/η)+1)(\ln (Q/\eta)+1) for Stochastic Submodular Cover: here QQ is the "goal value" and η\eta is the smallest non-zero marginal increase in utility deliverable by an item. (For integer-valued utility functions, we show a bound of H(Q)H(Q), where H(Q)H(Q) is the QthQ^{th} Harmonic number.) Although this bound was claimed by Golovin and Krause in the original version of their paper, the proof was later shown to be incorrect by Nan and Saligrama (2017). The subsequent corrected proof of Golovin and Krause (2017) gives a quadratic bound of (ln(Q/η)+1)2(\ln(Q/\eta) + 1)^2. Other previous bounds for the problem are 56(ln(Q/η)+1)56(\ln(Q/\eta) + 1), implied by work of Im et al. (2016) on a related problem, and k(ln(Q/η)+1)k(\ln (Q/\eta)+1), due to Deshpande et al. (2016) and Hellerstein and Kletenik (2018), where kk is the number of states. Our bound generalizes the well-known (ln m+1)(\ln~m + 1) approximation bound on the greedy algorithm for the classical Set Cover problem, where mm is the size of the ground set.

Keywords

Cite

@article{arxiv.2102.01149,
  title  = {A Tight Bound for Stochastic Submodular Cover},
  author = {Lisa Hellerstein and Devorah Kletenik and Srinivasan Parthasarathy},
  journal= {arXiv preprint arXiv:2102.01149},
  year   = {2021}
}

Comments

This work extends the result of Srinivasan Parthasarathy in his paper arXiv:1803.07639 from the problem of Stochastic Set Cover to that of Stochastic Submodular Cover

R2 v1 2026-06-23T22:44:31.684Z