Computational modelling of complex systems

Informatics is the engine of a new scientific revolution. The combination of big-data analysis and formal modelling is driving a new paradigm to study complex systems, from biology to economics. A central theme of this revolution is the network abstraction: a system is represented as a network of nodes performing local computations and interacting according to a graph structure. One important aspect of these systems is that their complexity unfolds in time: their emergent behaviour is the result of a dynamical process, with both their state and the structure of their interactions changing in time. Understanding how these system behave, and controlling them, is one of the grand scientific challenges of our times, which underlies many of the societal challenges of Horizon 2020, from personalised medicine to smart energy. Modelling such systems is difficult. Currently, there are many competing mathematical and computational approaches: applied mathematics, statistical physics, agent-based simulations, quantitative formal methods, machine learning. However, the interactions among these techniques are rather weak, leading to a fragmentation of methodologies that limits our current ability to understand, design, and control such a complexity. In this talk, I will present my research in a the perspective of an integrative approach, with quantitative formal methods (a branch of computer science born to study concurrent processes) playing the role of a bridge connecting different methodologies in a coherent framework. I will discuss the theory from a high-level perspective, emphasising the current and potential applications, encompassing areas as diverse as economics, biology, and smart cities.