Towards a Group Selection Design Pattern

Recent work has shown how socially inspired mechanisms developed within social science simulations can be applied to a number of application domains in distributed systems engineering. Here we proposed a general mechanism based on an evolutionary "group selection"" process in which individual entities (agents, nodes or components) dynamically copy the behavior of other entities if they achieve higher performance. We present this approach by beginning to sketch a generic "design pattern" and then show how it can be applied in different domains. Finally we discuss some open issues and future work.