Deriving exact results for Ising-like models from the cluster variation method

Abstract

The cluster variation method (CVM) is an approximation technique which generalizes the mean field approximation and has been widely applied in the last decades, mainly for finding accurate phase diagrams of Ising-like lattice models. Here we discuss in which cases the CVM can yield exact results, considering: (i) one-dimensional systems and strips (in which case the method reduces to the transfer matrix method), (ii) tree-like lattices and (iii) the so-called disorder points of euclidean lattice models with competitive interactions in more than one dimension.