Why are macroscopic experiments reproducible? Imitating the behavior of an ensemble by single pure states

Abstract

Evidently, physical experiments are practically reproducible even though the fully identical preparation of initial state wave functions is often far beyond experimental possibilities. It is thus natural to explore if and in which sense specific, uncontrollable features of initial wave functions are irrelevant for the observable course of an experiment. To this end we define ensembles of pure states which are then shown to generate extremely similar non-equilibrium dynamics of the expectation values of practically all standard observables. The ensembles are constructed to comply with some reduced, coarse a priori information on the state of the system, like, e.g. a few specific expectation values, etc. However, different types of ensembles with different additional properties are possible. We discuss some of them.