Accelerating Markov chain Monte Carlo simulation by differential evolution with self-adaptive randomized subspace sampling
- Los Alamos National Laboratory
- UNIV OF AMSTERDAM
Markov chain Monte Carlo (MCMC) methods have found widespread use in many fields of study to estimate the average properties of complex systems, and for posterior inference in a Bayesian framework. Existing theory and experiments prove convergence of well constructed MCMC schemes to the appropriate limiting distribution under a variety of different conditions. In practice, however this convergence is often observed to be disturbingly slow. This is frequently caused by an inappropriate selection of the proposal distribution used to generate trial moves in the Markov Chain. Here we show that significant improvements to the efficiency of MCMC simulation can be made by using a self-adaptive Differential Evolution learning strategy within a population-based evolutionary framework. This scheme, entitled DiffeRential Evolution Adaptive Metropolis or DREAM, runs multiple different chains simultaneously for global exploration, and automatically tunes the scale and orientation of the proposal distribution in randomized subspaces during the search. Ergodicity of the algorithm is proved, and various examples involving nonlinearity, high-dimensionality, and multimodality show that DREAM is generally superior to other adaptive MCMC sampling approaches. The DREAM scheme significantly enhances the applicability of MCMC simulation to complex, multi-modal search problems.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC52-06NA25396
- OSTI ID:
- 960766
- Report Number(s):
- LA-UR-08-07126; LA-UR-08-7126; TRN: US201008%%686
- Journal Information:
- International Journal of Nonlinear Sciences and Numerical Simulation, Journal Name: International Journal of Nonlinear Sciences and Numerical Simulation
- Country of Publication:
- United States
- Language:
- English
Similar Records
Bayesian calibration of terrestrial ecosystem models: a study of advanced Markov chain Monte Carlo methods
Application of Markov Chain Monte Carlo Methods for Uncertainty Quantification in Inverse Transport Problems