Dr. Xin Lin, Research Assistant in the Center, advised by Prof. Boleslaw Szymanski, defended his Ph.D. thesis titled "Preditcability Limit of Cascades in Interconnected Models".

Dr. Xin Lin, Research Assistant in the Center, advised by Prof. Boleslaw Szymanski, defended his Ph.D. thesis titled "Predictability Limit of Cascades in Interconnected Models". In the thesis, Dr. Lin analytically formulated abstract networks consisting of conceptual objects, such as the global risk network and the re propagation model. He also studied a real-world system, which is the European power grid. Dr. Lin proposed an innovative methodology to quantitatively analyze the cascading failures in a system consisting of multiple nonhomogeneous Poisson processes modeled as an Alternating Renewal Process (ARP) model of global risks in cross domains. He formulated a practical methodology to simulate complicated processes with hidden variables. To verify the parameter recovery using the method of maximum likelihood estimation (MLE), Dr. Lin compared the estimated parameters with ground-truth parameters in a similar model, which simulates the fi re propagation in a city and demonstrated that the convergence and asymptotic properties of the maximum likelihood estimation are consistent with the theoretical analysis.