modify monte_carlo criterion crt # 1 method meth # 2 dispose mem # 3 # archive arch # 4 update_list list # 5 print prin # 6 # acceptance acc # 7 step_size step # 8 adjust_step adj # 9 up_step up # 10 down_step down # 11 min_step limit # 12 end
Modify the Monte Carlo method. (1) The variable MTC_PASSED is set to true if the acceptance rate (proportion of attempted moves that are accepted) is within crt of the set target acc. (2) The only method available now is Metropolis, meth is preset to metropolis. (3) The Monte Carlo algorithm has somewhat different data requirements from the other simulation methods; thus each time this method is called, data structures are remapped to new ones and on completion of the simulation the additional data structures are removed. Setting mem to no will cause the data structures to be retained meaning that the remapping is not necessary on the second and subsequent invocations of the monte carlo method, the preset value of yes means that the data structures are remapped every time this method is used.
A cycle of monte carlo consists of as many attempted moves as there are number of atoms. The computation time for one cycle of monte carlo is approximately equivalent to the time for two molecular dynamics steps or about one energy minimization step. During a simulation (4) the coordinates are archived every arch cycles, preset value is 0 which means no archiving; (5) the non-bond lists are updated every list cycles, preset value 25 and (5) the energies are printed every prin cycles, preset value 1.
(7) The acceptance ratio acc, preset value 0.5, is a target for the proportion of attempted moves that meet the Boltzmann criterion for acceptance. (8) The simulation starts at a maximum step size of step, preset value 0.1 Å; each step consists of picking an atom at random and displacing it in a random direction for a random distance of 0 to step Å. (9) The maximum step size is adjusted every adj steps, preset value 1, by scaling by (10) the factor up (preset value 1.05) if the current acceptance ratio exceeds the target or by (11) the factor down (preset value 0.95) if the current acceptance ratio is below the target acc. (12) This setting keeps the step size setting from falling below limit. The preset value is 0.01Å. Under some circumstances, the step size may be adjusted to a zero value at which point it can no longer be increased. The simulation will come to a stop.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|