Nonequilibrium systems



The statistical concept of entropy is the basis for related criteria

of stability and minimum-bias estimation.  Though first shown useful

for equilibrium physical systems, its extensive development in the

theory of communication and computation has shown that entropy methods

are inherently methods of inference, not restricted to any particular

domain or substrate.  In principle, entropies can be defined for

non-equilibrium quantum systems, ensembles of histories, and even

evolutionary dynamics -- anywhere a system's internal degrees of

freedom are not fully determined by the constraints on its boundaries

or the laws of its dynamics. 

Opening statistical inference to the domain of processes as well as

states leads to the idea of "kinetically constrained ensembles".

Single chemical reactions constrained by quantum bond structure, whole

reaction networks constrained by stoichiometry and reaction rates, and

especially the multi-layered spatial and catalytic systems of

biochemistry are examples of ensembles with kinetic as well as

energetic constraints.  Such ensembles can be described

macroscopically by state variables which are currents as well as by

configurations.  New kinds of phase transitions can occur in which

fluxes are the order parameters.  If the emergence of life took place

through a series of dynamical phase transitions in driven

geochemistry, non-equilibrium entropy methods may provide a way to

explain the emergence and subsequent stability of the biosphere.



Supriya Krishnamurthy

Cosma Shalizi

Steve Frank

Walter Fontana

David Krakauer