Accelerating Markov Chain Monte Carlo Simulation by Differential Evolution with Self-Adaptive Randomized Subspace Sampling

By Jasper A. Vrugt, C. T. ter Braak, Cees G. H. Diks, Bruce A. Robinson, James M. Hyman, D. Higdon. 2009

1,086 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/627f7b0c308e05be60b71d537db60116db5b3fdd

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Confronting uncertainty in model-based geostatistics using Markov Chain Monte Carlo simulation (2011, methodology): “A detailed algorithmic description of DREAM appears in Vrugt et al. (2008) and Vrugt et al. (2009) and so will not be repeated here.”
Training‐Image Based Geostatistical Inversion Using a Spatial Generative Adversarial Neural Network (2017, methodology): “…are compared the As no analytical solution of p ( θ | d ) is available for the type of non-linear inverse problems considered herein, we sample from p ( θ | d ) by MCMC simulation (see, e.g., Robert and Casella , 2004) with the DREAM (ZS) algorithm ( Vrugt et al. , 2009; Laloy and Vrugt , 2012).”
Evaluating marginal likelihood with thermodynamic integration method and comparison with several other numerical methods (2016, methodology): “…numerical comparison, it is realized that the Metropolis-Hasting MCMC techniques used in Elsheikh et al. [2013] for implementing nested sampling are standard and do not have advanced features of the DREAM code [Vrugt et al., 2009; Laloy and Vrugt, 2012] used to implement thermodynamic integration.”
Parameter Analysis and Estimates for the MODIS Evapotranspiration Algorithm and Multiscale Verification (2019, methodology): “…Bayesian‐based methods, such as Markov‐chain Monte Carlo (MCMC) analysis, often suffer from problems related to selection of an appropriate scale and orientation for the jumping distribution, which may prevent the algorithm from efficiently converging (Haario et al., 2006; Vrugt et al., 2009).”
Corruption of accuracy and efficiency of Markov chain Monte Carlo simulation by inaccurate numerical implementation of conceptual hydrologic models (2010, methodology): “[23] In this paper, we use the recently developed DiffeRential Evolution Adaptive Metropolis (DREAM) algorithm [Vrugt et al., 2009] to generate samples from the posterior distribution.”
Global hydrology 2015: State, trends, and directions (2015, positive): “Whole generations of hydrologists have been focusing on parameter nonuniqueness or equifinality [Beven, 1993] and developed sophisticated and efficient Markov Chain Monte Carlo methods to deal with the parameter estimation under equifinality [e.g., Beven and Binley, 1992; Duan et al., 1992; Vrugt et”
Assessment of parametric uncertainty for groundwater reactive transport modeling (2014, methodology): “DREAM(ZS) uses Differential Evolution Adaptive Metropolis (DREAM) [Vrugt et al., 2008a, 2009a, 2009b] as its main building block, which runs multiple Markov chains from different starting points in parallel for global search in the parameter space to make it possible to search multiple parameter…”
Using model averaging to combine soil property rasters from legacy soil maps and from point data (2014, positive): “The DREAM algorithm implemented for BMA (Vrugt et al., 2009) was run using script developed for Matlab (Mathworks 2012).”
On-the-Job Search and Precautionary Savings (2013, methodology): “Details pertaining to tuning the MCMC algorithm, a parallel implementation, and related methods in statistics can be found in Robert and Casella (2004), Vrugt et al. (2009), and Sisson and Fan (2011).”
Ancient numerical daemons of conceptual hydrological modeling: 2. Impact of time stepping schemes on model analysis and prediction (2010, methodology): “Since current practice is progressively dominated by Markov chain Monte Carlo (MCMC) methods [e.g., Kuczera and Parent, 1998; Vrugt et al., 2009], we also provide some comments on the effect of time stepping artifacts on the convergence of MCMC algorithms.”
A formal likelihood function for parameter and predictive inference of hydrologic models with correlated, heteroscedastic, and non‐Gaussian errors (2010, methodology): “DREAM‐ZS is based on the original DREAM algorithm [Vrugt et al., 2009] but uses sampling from an archive of past states to generate candidate points in each individual chain.”