Figure 2

Flowchart representing the basic scheme of the out-of-equilibrium adaptive Monte Carlo (OEAMC) algorithm. Given an initial parameter point θ₀, covariance matrix ∑ and β, the algorithm carries out n iterations in which every new parameter point is sampled from a normal distribution (4), and accepted or rejected based on Metropolis acceptances ratios (5). Every n iterations the viable points (blue and black points in the figure correspond to viable and nonviable sampled parameter points, respectively) found so far are grouped into clusters and the volume (grey areas in the figure) of ellipsoids that enclose the viable parameter points in each cluster is calculated. If the sum of these volumes converges the algorithm stops; if not, the covariance matrix ∑ and β are updated (6), and n new iterations are performed. The output of the algorithm is the set V _MC which includes all the viable parameter points found.