Distributed Strategy Selection: A Submodular Set Function Maximization Approach