Importance sampling large deviations in nonequilibrium steady states. I

Our first paper with nonequilbrium steady states used three systems to illustrate the sampling issues that arise in these systems. We contrast two techniques - the Cloning Algorithm (CA) and Transition Path Sampling (TPS) and show that the fundamental problem of exponentially growing variance is present in both techniques. Crucially, we show that it is possible to extract the rate function as well as the cumulant generating function using both TPS and CA. We propose an alternate weighting scheme to compute non-convex rate functions efficiently and demonstrate its applicability to self-assembling systems.