English

Fast algorithms for k-submodular maximization subject to a matroid constraint

Data Structures and Algorithms 2023-07-27 v1 Machine Learning

Abstract

In this paper, we apply a Threshold-Decreasing Algorithm to maximize kk-submodular functions under a matroid constraint, which reduces the query complexity of the algorithm compared to the greedy algorithm with little loss in approximation ratio. We give a (12ϵ)(\frac{1}{2} - \epsilon)-approximation algorithm for monotone kk-submodular function maximization, and a (13ϵ)(\frac{1}{3} - \epsilon)-approximation algorithm for non-monotone case, with complexity O(n(kEO+IO)ϵlogrϵ)O(\frac{n(k\cdot EO + IO)}{\epsilon} \log \frac{r}{\epsilon}), where rr denotes the rank of the matroid, and IO,EOIO, EO denote the number of oracles to evaluate whether a subset is an independent set and to compute the function value of ff, respectively. Since the constraint of total size can be looked as a special matroid, called uniform matroid, then we present the fast algorithm for maximizing kk-submodular functions subject to a total size constraint as corollaries. corollaries.

Keywords

Cite

@article{arxiv.2307.13996,
  title  = {Fast algorithms for k-submodular maximization subject to a matroid constraint},
  author = {Shuxian Niu and Qian Liu and Yang Zhou and Min Li},
  journal= {arXiv preprint arXiv:2307.13996},
  year   = {2023}
}
R2 v1 2026-06-28T11:40:22.498Z