dili

Dimension-independent likelihood-informed MCMC

Topics

bayesian_calibration

Specification

• Alias: None

• Arguments: None

• Default: dram

Child Keywords:

Required/Optional	Description of Group	Dakota Keyword	Dakota Keyword Description
Optional		hessian_type	Parameter for the DILI MCMC algorithm in MUQ
Optional		adapt_interval	Parameter for the DILI MCMC algorithm in MUQ
Optional		adapt_start	Parameter for the DILI MCMC algorithm in MUQ
Optional		adapt_end	Parameter for the DILI MCMC algorithm in MUQ
Optional		initial_weight	Parameter for the DILI MCMC algorithm in MUQ
Optional		hess_tolerance	Parameter for the DILI MCMC algorithm in MUQ
Optional		lis_tolerance	Parameter for the DILI MCMC algorithm in MUQ
Optional		ses_num_eigs	Parameter for the stochastic eigensolver used by the DILI MCMC algorithm in MUQ
Optional		ses_rel_tol	Parameter for the stochastic eigensolver used by the DILI MCMC algorithm in MUQ
Optional		ses_abs_tol	Parameter for the stochastic eigensolver used by the DILI MCMC algorithm in MUQ
Optional		ses_exp_rank	Parameter for the stochastic eigensolver used by the DILI MCMC algorithm in MUQ
Optional		ses_overs_factor	Parameter for the stochastic eigensolver used by the DILI MCMC algorithm in MUQ
Optional		ses_block_size	Parameter for the stochastic eigensolver used by the DILI MCMC algorithm in MUQ

Description

When the user selects ‘bayesian_calibration muq’ in the Dakota input file, the occurrence of the ‘dili’ keyword will cause Dakota to use the DILI algorithm ([CLM16]) from the MUQ library.