Efficient use of Monte Carlo simulation data
Efficient techniques for using Monte Carlo simulation data are presented. These techniques make use of the probability distribution (histogram) generated by the simulation to study the behavior of the system over a range of temperatures, fields, it etc. The first of these techniques, the single-histogram method, was introduced thirty years ago but went largely unnoticed as a powerful tool in the study of phase transitions. In this thesis it is shown that by using this technique, a single simulation near a phase transition is sufficient to provide information over the entire finite-size scaling region. This is demonstrated by studies of the d = 2 Ising model and the d = 2 8-state Potts model. The second technique, the multiple-histogram method, is a new method for combining the results of more than one Monte Carlo simulation. This method provides an optimized combination of data from different sources, produces results in the form of continuous functions and can be applied to an arbitrary number of simulations. Errors can be calculated and provided a clear and simple guide to optimizing the length and location of additional simulations to provide maximum accuracy. The method is tested on the d = 2 Ising model.
- Research Organization:
- Carnegie-Mellon Univ., Pittsburgh, PA (USA)
- OSTI ID:
- 5925889
- Country of Publication:
- United States
- Language:
- English
Similar Records
Monte Carlo renormalization group
Two-sided bounds on the free energy from local states in Monte Carlo simulations