We investigate a non-submodular maximization problem subject to a p-independence system constraint, where the non-submodularity of the utility function is characterized by a series of parameters, such as submodularity (supmodularity) ratio, generalized curvature, and zero order approximate submodularity coefficient, etc. Inspired by Feldman et al. [15] who consider a non-monotone submodular maximization with a p-independence system constraint, we extend their Repeat-Greedy algorithm to non-submodular setting. While there is no general reduction to convert algorithms for submodular optimization problems to non-submodular optimization problems, we are able to show the extended Repeat-Greedy algorithm has an almost constant approximation ratio for non-monotone non-submodular maximization.
CITATION STYLE
Yang, R., Xu, D., Du, D., Xu, Y., & Yan, X. (2019). Maximization of Constrained Non-submodular Functions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11653 LNCS, pp. 615–626). Springer Verlag. https://doi.org/10.1007/978-3-030-26176-4_51
Mendeley helps you to discover research relevant for your work.