Improved randomized algorithm for $k$-submodular function maximization

Abstract: Submodularity is one of the most important properties in combinatorial optimization, and $k$-submodularity is a generalization of submodularity. Maximization of a $k$-submodular function requires an exponential number of value oracle queries, and approximation algorithms have been studied. For unconstrained $k$-submodular maximization, Iwata et al. gave randomized $k/(2k-1)$-approximation algorithm for monotone functions, and randomized $1/2$-approximation algorithm for nonmonotone functions. In this paper, we present improved randomized algorithms for nonmonotone functions. Our algorithm gives $\frac{k^2+1}{2k^2+1}$-approximation for $k\geq 3$. We also give a randomized $\frac{\sqrt{17}-3}{2}$-approximation algorithm for $k=3$. We use the same framework used in Iwata et al. and Ward and \v{Z}ivn\'{y} with different probabilities.