A METROPOLIS MONTE-CARLO METHOD FOR COMPUTING MICROCANONICAL STATISTICAL RATE CONSTANTS

A new method of computing microcanonical statistical rate constants is presented. The method utilizes the Metropolis Monte Carlo algorithm in a manner which circumvents some of the numerical inefficiency associated with other Metropolis and "shot-gun" Monte Carlo based schemes. It is therefore expected to be useful in studies of many degree of freedom systems where numerical efficiency is crucial. Optimization of the method efficiency with respect to its adjustable parameters is examined in detail, both theoretically and in a numerical study of the T-shaped Ar3 "inversion" process. The energy dependence of the T-shaped Ar3 inversion rate is studied in a sample application of the method. An application to full three dimensional Ar3 will be presented in a future study.