We focus on three research thrusts representing key challenges and opportunities in the field of coupled human-natural systems.

Controlling interconnected infrastructures. The dynamic interconnection between multiple infrastructure systems and socio-economic sectors challenges us with many problems. Does the interaction between different systems worsen or lessen the impact of extreme events? Are there management solutions we fail to identify when focusing on individual sectors? Traditional simulation analysis and operations research tools are ill-suited to fully grasp the complexity of these challenges. In fact, there is a growing chasm between the models simulating the dynamics of coupled human-natural systems and the optimization methods supporting daily decision-making, because the latter are often constrained by specific problem formulations. To bridge this gap, we work on algorithms and software for the co-simulation and control of interconnected infrastructure systems.

Developing hybrid data-driven and process-based models. A fundamental challenge in the adoption of process-based simulation (and co-simulation) models stems from their computational requirements, which often shadow the rapid increase in computing power—even when parallelization and large-scale clusters are adopted. When developing a model, we therefore face a tradeoff between its capacity to realistically represent the system of interest and its capacity to solve computationally-intensive problems, such as real-time control or risk analysis. How do we resolve this epistemic deadlock? How do we scale our solutions to large domains? To answer these questions, we work on the synergistic combination of data science and process-based modelling.