Monte Carlo The Classic Algorithm

a trial move is made to generate a new trial configuration, according to a rule that we call a move strategy. The simple move strategy introduced by Metropolis et al. is still the most widely used method. One 

Translate an atom within a. small cube a random amount 

Pick a space-fixed axis at random rigidly rotate body about this axis by a random amount 

The parameter L may have an optimum value for the particular type of atom of interest, as well as for the temperature and other variables studied. If L is too small, most moves will be accepted and a very large number of attempts will be required to move very far from the initial configuration. Flowever, if L is too big very few trial moves will be accepted and again, the walker will require many steps to move away from the starting point. For this reason one generally chooses L so that between 30% and 70% of the moves are accepted (50% is a happy medium). Each atom can be moved in sequence. 

Similar formulas hold for U ) and U ) . It is important to emphasize that configurations are not excluded for averaging purposes simply because they have been most recently accepted or rejected. Both accepted and rejected configurations must be included in the average, or the potential energies will not be Boltzmann-distributed. 

---