Complex systems are often represented as networks of interacting objects. The way these networks are internally organised is often non-trivial and includes fairly independent sub-structures (e.g social groups, news articles on the same topic, biochemical pathways). Several random models have been proposed to study the networks topology, but not many of them are able to generate real-world like networks.
In this talk, I will present an overview of most popular random graph models and discuss why they are not well suited for the complexity of the real-world. Then I will describe our own model that can generate complex and realistic topology with community structure and demonstrate how to tune its parameters with an evolution inspired optimisation algorithm (CMA-ES).