ABSTRACT
Algorithmic randomness is defined without a recourse to probabilities. It provides an alternative to the usual ensemble measures of disorder: it quantifies randomness of the known features of the state of the physical system. I shall demonstrate that it is indispensable in formulating thermodynamics from the viewpoint of the information gathering and using system ("IGUS")—a Maxwell's demon-like entity capable of performing measurements and of modifying its strategies (for example, for extraction of useful work) on the basis of the outcomes of the measurements. Such an IGUS can be regarded as a "complex adaptive system." The aim of this paper is to review the concept of the algorithmic information content in the context of statistical mechanics and discuss its recently discovered physical applications.
