Binary tree summation Monte Carlo method for Potts models
We give a new sampling algorithm for the Potts model based on the Fortuin-Kasteleyn transformation. The method produces independent samples and sums up a large number of configurations for each sweep. The partition function and thermodynamic averages for all values of the temperature can be computed from a single run. We compare the results with exact 2D Ising model.