Configurational bias Monte Carlo: Difference between revisions
(New page: It is usual that many of the accessible configurations have a small probability and only a few ones are probable. In these cases, the simulation is more efficient if the probabilities of t...) |
No edit summary |
||
| Line 1: | Line 1: | ||
It is usual that many of the accessible configurations have a small probability and only a few ones are probable. In these cases, the simulation is more efficient if the probabilities of the different configurations are previously considered. With this end, the new position for a unit is randomly chosen between a discrete number of possibilities (the neighboring sites in lattice models or a randomly chosen set of positions in other cases), taking into account their Boltzmann probabilities. In the case of polymers, an entirely new part of a chain up to an end can be generated following a path of easily accessible positions. This introduces a bias which should be compensated by considering a weight factor for each new position chosen (or a product of these factors for a new chain). A similar weight corresponding to reconstructing the old configuration from the new one has also to be calculated. The probability ratios are corrected by introducing the ratio between the new and the old configurational weight factors. | It is usual that many of the accessible configurations have a small probability and only a few ones are probable. In these cases, the simulation is more efficient if the probabilities of the different configurations are previously considered. With this end, the new position for a unit is randomly chosen between a discrete number of possibilities (the neighboring sites in lattice models or a randomly chosen set of positions in other cases), taking into account their Boltzmann probabilities. In the case of polymers, an entirely new part of a chain up to an end can be generated following a path of easily accessible positions. This introduces a bias which should be compensated by considering a weight factor for each new position chosen (or a product of these factors for a new chain). A similar weight corresponding to reconstructing the old configuration from the new one has also to be calculated. The probability ratios are corrected by introducing the ratio between the new and the old configurational weight factors. | ||
I. Siepman and D. Frenkel, Mol. Phys. 75, 59 1992 | |||
==References== | |||
1. I. Siepman and D. Frenkel, Mol. Phys. 75, 59 (1992). | |||
Revision as of 18:16, 13 June 2007
It is usual that many of the accessible configurations have a small probability and only a few ones are probable. In these cases, the simulation is more efficient if the probabilities of the different configurations are previously considered. With this end, the new position for a unit is randomly chosen between a discrete number of possibilities (the neighboring sites in lattice models or a randomly chosen set of positions in other cases), taking into account their Boltzmann probabilities. In the case of polymers, an entirely new part of a chain up to an end can be generated following a path of easily accessible positions. This introduces a bias which should be compensated by considering a weight factor for each new position chosen (or a product of these factors for a new chain). A similar weight corresponding to reconstructing the old configuration from the new one has also to be calculated. The probability ratios are corrected by introducing the ratio between the new and the old configurational weight factors.
References
1. I. Siepman and D. Frenkel, Mol. Phys. 75, 59 (1992).