Alternatives to conventional ensemble averages for thermodynamic properties

Long-established molecular theories often can provide good analytical estimates of thermodynamic properties for simple molecular models. Molecular simulation, on the other hand, can provide exact properties for the same models, or others that are arbitrarily realistic, but requires significantly more computational expense. Recent developments enable molecular simulation to improve performance by exploiting information available from approximate theories. The connection is formed via alternative ensemble averages, which allow direct calculation of the correction to the theory. In favorable cases this correction is small and can be computed with low uncertainty. Alternative ensemble averages of this type can be formulated systematically using a recently introduced framework, known as mapped averaging.