The random cluster process (Q1344626)

Given a graph \(G\) and parameters \(0 < p < 1\) and \(Q > 0\), the general random cluster model gives a distribution on the edges of \(G\). The model was introduced by \textit{C. M. Fortuin} and \textit{B. W. Kasteleyn} in 1972. When \(Q = 1\), it is usual random subgraph model, in which edges of \(G\) are kept independently with probability \(p\). For \(Q \neq 1\) there are correlations. The model contains the Ising model and the \(Q\)-state Potts model. The present paper discusses the relationship of the model to several areas more familiar to combinatorialists, in particular the Tutte polynomial, reliability, electrical networks and percolation. Several identities and natural monotonicities are established, and open problems are presented.