Archive for May 30th, 2011

Detecting early-warning signals in complex dynamical systems

Monday, May 30th, 2011 by Antonio Manzalini

I’ve concluded my last post (“Can Quantum Mechanics contribute to Social Sciences ?”) with a question: can we early detect tipping points at which sudden regime transition of a complex system’s behaviour may happen ?

Predicting such critical points before they are reached is extremely difficult, but it might have a huge impact in several fields, from medicine to business, from meteorology to social networking, from new materials to control of future networks.

Please have a look at this paper “Early-warning signals for critical transitions”

http://www.nature.com/nature/journal/v461/n7260/abs/nature08227.html

Paper is suggesting the analysis of generic early-warning signals indicating, for a wide class of complex systems, the approaching of a critical threshold, where small forces can cause major changes in the state. Examples of such transitions might include the collapse of over-harvested ecosystems, climatic changes, or stocks markets dynamics.

For example one symptom is the critical slowing down: when the system approaches a critical transition, it becomes increasingly slow in recovering from small perturbations (which is translated mathematically into an increase in the autocorrelation and variance of the fluctuations).

Another signal that can be seen in the vicinity of a catastrophic transition point is flickering. Stochastically, the system moves back and forth between the basins of attraction of two alternative attractors (bistable region).

Spatial patterns is a third example: an ecosystem may show a predictable sequence of self-organized spatial patterns as it approaches a critical transition (e.g. a semi-arid vegetation to increasing dryness of the climate).

Still many challenges have to be overcome: for example, how collecting meaningful data, filtering techniques to increase the sensitivity of indicators, preventing false positives and others.

In any case progresses in detecting early-warning signals might be valuable in several fields. I wonder if this theory can also be used in controlling stability of future networks and related business ecosystems.