14
31

Multi-Objective Maximization of Monotone Submodular Functions with Cardinality Constraint

Abstract

We consider the problem of multi-objective maximization of monotone submodular functions subject to cardinality constraint, often formulated as maxA=kmini{1,,m}fi(A)\max_{|A|=k}\min_{i\in\{1,\dots,m\}}f_i(A). While it is widely known that greedy methods work well for a single objective, the problem becomes much harder with multiple objectives. In fact, Krause et al.\ (2008) showed that when the number of objectives mm grows as the cardinality kk i.e., m=Ω(k)m=\Omega(k), the problem is inapproximable (unless P=NPP=NP). On the other hand, when mm is constant Chekuri et al.\ (2010) showed a randomized (11/e)ϵ(1-1/e)-\epsilon approximation with runtime (number of queries to function oracle) nm/ϵ3n^{m/\epsilon^3}. %In fact, the result of Chekuri et al.\ (2010) is for the far more general case of matroid constant. We focus on finding a fast and practical algorithm that has (asymptotic) approximation guarantees even when mm is super constant. We first modify the algorithm of Chekuri et al.\ (2010) to achieve a (11/e)(1-1/e) approximation for m=o(klog3k)m=o(\frac{k}{\log^3 k}). This demonstrates a steep transition from constant factor approximability to inapproximability around m=Ω(k)m=\Omega(k). Then using Multiplicative-Weight-Updates (MWU), we find a much faster O~(n/δ3)\tilde{O}(n/\delta^3) time asymptotic (11/e)2δ(1-1/e)^2-\delta approximation. While the above results are all randomized, we also give a simple deterministic (11/e)ϵ(1-1/e)-\epsilon approximation with runtime knm/ϵ4kn^{m/\epsilon^4}. Finally, we run synthetic experiments using Kronecker graphs and find that our MWU inspired heuristic outperforms existing heuristics.

View on arXiv
Comments on this paper