Raissa D'Souza – UC Davis
Wednesday, April 23rd , 10:45AM DST
Self-organized criticality on networks: from black swans to dragon kings
The paradigm of self-organized criticality is a celebrated concept in statistical physics. A classic example is the Bak-Tang-Wiesenfeld (BTW) sandpile model on a two-dimensional lattice, which displays a power-law distribution of cascade sizes. Large events in such distributions have colloquially become referred to as “black swans”. Instead of lattices, interdependent networks are at the core of modern society, so what happens to the BTW model on a network? First, the power law behavior is largely robust. But more importantly, it will be shown in this talk how the BTW model can be used to study properties of networks like interdependence and systemic risk. The BTW model has also been used to model cascades in brain networks and electric power grids although it neglects the oscillatory nature of the nodes. By coupling the BTW model to the quintessential model of synchronization, the Kuramoto model, we show that this gives rise to a new range of emergent phenomena. In particular, the occurrence of self-amplifying cascades that lead to periodic system-wide failure that can be considered “dragon kings”. Also shown is that such dragon king events can exist even in the classic BTW model and may explain the large bumps seen in previous work that could not be attributed to finite size effects. We discuss the relation to self-organization around first-order and second-order absorbing state phase transitions.