| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 9727565 | Physica A: Statistical Mechanics and its Applications | 2005 | 13 Pages |
Abstract
In this paper, we propose to estimate the normalizing constants for simulated tempering by a modified histogram algorithm, the so-called contour Monte Carlo algorithm, and compare the efficiency of simulated tempering and parallel tempering. Our analysis reveals that simulated tempering tends to mix faster than parallel tempering at low temperature levels for simulating from complex systems. The reason why simulated tempering is better than parallel tempering is discussed at length.
Related Topics
Physical Sciences and Engineering
Mathematics
Mathematical Physics
Authors
Faming Liang,