Distributed Maximization of Submodular and Approximately Submodular Functions

TL;DR

This paper introduces a distributed greedy algorithm for maximizing submodular and approximately submodular functions across multiple agents with limited communication, achieving near-optimal solutions with performance guarantees.

Contribution

It presents a novel distributed approach for submodular maximization that balances communication complexity and solution quality, extending to approximate submodularity.

Findings

The algorithm converges to a (1-1/e) approximation of the optimal solution.

Performance depends on the number of communication steps.

The approach extends to non-submodular functions using approximate submodularity.

Abstract

We study the problem of maximizing a submodular function, subject to a cardinality constraint, with a set of agents communicating over a connected graph. We propose a distributed greedy algorithm that allows all the agents to converge to a near-optimal solution to the global maximization problem using only local information and communication with neighbors in the graph. The near-optimal solution approaches the (1-1/e) approximation of the optimal solution to the global maximization problem with an additive factor that depends on the number of communication steps in the algorithm. We then analyze convergence guarantees of the proposed algorithm. This analysis reveals a tradeoff between the number of communication steps and the performance of the algorithm. Finally, we extend our analysis to nonsubmodular settings, using the notion of approximate submodularity.