On the sampling requirements for exponential-work free-energy calculations

Calculation of free-energy differences using exponential-work averaging methods is examined, focusing on the amount of sampling required to converge to an accurate result. Prior research that has demonstrated that the necessary amount of sampling scales exponentially with the dissipated work associated with the reverse trajectory is reviewed. The result is demonstrated through a pedagogical application to a model process involving adiabatic compression and expansion of a gas.