'Expansion entropy': A new litmus test for chaos?

July 28, 2015, University of Maryland
Researchers from the University of Maryland have proposed a new definition of chaos that applies to a wide variety of chaotic systems, including attractors, (top left, a "Tinkerbell" attractor), repellers, (right side, top and bottom, illustrating chaotic scattering), and forced systems (bottom left, motions of a damped pendulum). The bottom right picture shows the pattern of light created inside a pyramid of four reflective balls. Credit: UMD Chaos Group

Can the flap of a butterfly's wings in Brazil set off a tornado in Texas? This intriguing hypothetical scenario, commonly called "the butterfly effect," has come to embody the popular conception of a chaotic system, in which a small difference in initial conditions will cascade toward a vastly different outcome in the future.

Understanding and modeling chaos can help address a variety of scientific and engineering questions, and so researchers have worked to develop better mathematical definitions of chaos. These definitions, in turn, will aid the construction of models that more accurately represent real-world chaotic systems.

Now, researchers from the University of Maryland have described a new definition of chaos that applies more broadly than previous definitions. This new definition is compact, can be easily approximated by numerical methods and works for a wide variety of chaotic systems. The discovery could one day help advance computer modeling across a wide variety of disciplines, from medicine to meteorology and beyond. The researchers present their new definition in the July 28, 2015 issue of the journal Chaos.

"Our definition of chaos identifies even when it lurks in the dark corners of a model," said Brian Hunt, a professor of mathematics with a joint appointment in the Institute for Physical Science and Technology (IPST) at UMD. Hunt co-authored the paper with Edward Ott, a Distinguished University Professor of Physics and Electrical and Computer Engineering with a joint appointment in the Institute for Research in Electronics and Applied Physics (IREAP) at UMD.

The study of chaos is relatively young. MIT meteorologist Edward Lorenz, whose work gave rise to the term "the butterfly effect," first noticed chaotic characteristics in weather models in the mid-20th century. In 1963, he published a set of differential equations to describe atmospheric airflow and noted that tiny variations in initial conditions could drastically alter the solution to the equations over time, making it difficult to predict the weather in the long term.

Mathematically, extreme sensitivity to initial conditions can be represented by a quantity called a Lyapunov exponent. This number is positive if two infinitesimally close starting points diverge exponentially as time progresses. Yet, Lyapunov exponents have limitations as a definition of chaos: they only test for chaos in particular solutions of a model, not in the model itself, and they can be positive even when the underlying model is considered too straightforward to be deemed chaotic.

Fittingly, the chaotic solution to Lorenz's equations looks like two wings of a butterfly. The shape can be mathematically categorized as an attractor, meaning it is easy to identify with Lyapunov exponents. Yet, not all chaotic behavior is quite so clear, Hunt explained.

To provide an example, Hunt described four glass-ball Christmas ornaments stacked in a pyramid. Light hitting the shiny spheres reflects off in all directions. Most of the light travels along simple paths, but some photons can become trapped in the interior of the pyramid, bouncing chaotically back and forth between the ornaments. These chaotic light paths are mathematically categorized as repellers and can be difficult to find using model equations unless you know exactly where to look.

Researchers commonly encounter chaotic repellers in physical and natural systems as varied as plumbing networks, asteroid orbits, chemical reactions, geophysical systems, bird flocks and human organ systems.

To fit the generally recognized forms of chaos under one umbrella definition, Hunt and Ott turned to a concept called entropy. In a system that changes over time, entropy represents the rate at which disorder and uncertainty build up.

The idea that entropy could be a proxy for chaos is not new, but the standard definitions of entropy, such as metric entropy and topological entropy, are trapped in the mathematical equivalent of a straightjacket. The definitions are difficult to apply computationally and contain stringent prerequisites that make it difficult or impossible to apply the definitions to many physical and biological systems of interest to scientists.

Hunt and Ott's work defines a new flexible type of entropy, called expansion entropy, which can be applied to more realistic models of the world. The researchers also broadened their definition of chaos by including systems where external factors continue to push or pull on the model as it evolves. The definition can be approximated accurately by a computer and can accommodate systems, like regional weather models, which are forced by potentially chaotic inputs. The researchers define chaotic models as those that exhibit positive expansion entropy.

The researchers hope expansion entropy will become a simple, commonly used tool to identify chaos in a wide range of model systems. Pinpointing chaos in a system can be a first step to determining whether the system can ultimately be controlled.

For example, Hunt explains, two identical chaotic systems with different initial conditions may evolve completely differently, but if the systems are forced by external inputs they may start to synchronize. By applying the expansion entropy definition of chaos and characterizing whether the original systems respond chaotically to inputs, researchers can tell whether they can wrestle some control over the through inputs to the system.

This type of control could be used, for example, to design highly secure communications systems and more effective pacemakers for the heart, Hunt said.

Explore further: Chaos is an inherent part of city traffic

More information: The research paper, "Defining Chaos," Brian Hunt and Edward Ott, was published on July 28, 2015 in the journal Chaos: An Interdisciplinary Journal of Nonlinear Science: DOI: 10.1063/1.4922973

Related Stories

Chaos is an inherent part of city traffic

July 21, 2015

It's not unusual for two drivers to depart from the same location, head out to the same destination, drive more or less the same speed and nevertheless arrive at dramatically different times, with one driver taking significantly ...

Probing the edge of chaos

February 27, 2014

The edge of chaos—right before chaos sets in—is a unique place. It is found in many dynamical systems that cross the boundary between a well-behaved dynamics and a chaotic one. Now, physicists have shown that the distribution—or ...

Mathematics confirm the chaos of the Spanish labor market

January 25, 2012

Unemployment time series in Spain behave in a chaotic way according to a study at the University of Seville. Such chaos demonstrates the complex and unpredictable nature of the Spanish labour market in the long run. However, ...

Recommended for you

Topological quantities flow

January 14, 2019

Topology is an emerging field within many scientific disciplines, even leading to a Nobel Physics Prize in 2016. Leiden physicist Marcello Caio and his colleagues have now discovered the existence of topological currents ...

3 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

TimLong2001
not rated yet Jul 29, 2015
Nonlinear ("chaos") theory is deterministic. The dynamics of the photon dictate gradual energy loss, leading to a red shift sans expansion.
Hyperfuzzy
not rated yet Jul 30, 2015
No, a butterfly does not enhance or cause chaos. The fundamentals applied correctly, i.e. the state of the particle distribution is never chaotic. The system always seeks a stable state and will self assemble. One must define the space for the chaotic event as a contained state within a confined region without added energy or matter. No, chaos is only a mathematical state. That is, turbulence will always decrease without added input. Therefore, the minor input relative to the earth of a butterfly would not follow physics, unless its a magical butterfly that may select the modal response of its wings to create said chaos. A very difficult calculation, but not chaotic, controlled chaos.
Hyperfuzzy
not rated yet Jul 30, 2015
The photon, or rather a point upon the sphere of the object of transmission proportional intensity as 1/surface-area of the sphere of the charge, else its a defined point within the spot size of a beam from a controlled source. Not sure why this would relate to chaos? The red shift is a function of the transmitter, TBD, and the relative velocity of the receiver. Not sure what nonlinear chaos is relative to chaos but matter has a body functional response ...

Note the field intensity is defined by charge motion added to the static state. This state is defined at a single instant in time. However, the 4D space defines a continuum, just define it as unit-less, map the unit from c=lambda nu, then choose either length or time, with all dimensions of the same unit. remove the units then and we define an isomorphic space to reality multiplier of unit vector i4 is zero when taking this surface area.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.