Distributed strategy selection: A submodular set function maximization approach

Joint utility-maximization problems for multi-agent systems often should be addressed by distributed strategy-selection formulation. Constrained discrete feasible strategy sets, these are broadly formulated as NP-hard combinatorial optimization problems. In many cases, can cast constrained submodular set function maximization problems, which also belong to the domain of A prominent example is problem mobile sensor dispatching over a domain. This paper considers class that consist monotone and subject partition matroid constraint group networked agents communicate connected undirected graph. We work with value oracle model. Consequently, only access utility through black box returns given specific set. propose suboptimal polynomial-time algorithm enables each agent obtain its respective via local interactions neighboring agents. Our solution fully gradient-based using functions’ multilinear extension followed stochastic Pipage rounding procedure. results in when team evaluated at worst case, 1c(1−e−c−O(1/T)) optimal c being curvature function. An demonstrates our results.