On Population-Based Algorithms for Distributed Constraint Optimization Problems