1Blurb:: 2Variance applied to simulation responses 3 4Description:: 5The variance that is applied to simulations run by Dakota, i.e. using the 6\c analysis_drivers command. The user may supply a single variance 7or a vector of variances of length equal to the number of responses. In both 8cases, the values provided are treated as scalar variance types. If a single 9variance is provided, it is applied to all responses produced by the 10simulation code. If a vector is provided, each variance is applied to the 11corresponding response output produced by the simulation code. 12 13It is important to note that the the variance defined by this keyword differs 14from that defined using \c experiment_variance_type. These two commands apply 15to user-provided calibration data, specified, for example, by 16\c calibration_data or \c calibration_data_file. However, \c simulation_variance 17applies to those responses produced by simulation code that is run by Dakota, 18as described above. 19 20<b> Usage Tips </b> 21 22Currently, this keyword is only in use as part of the algorithm implemented by 23\ref method-bayes_calibration-experimental_design. In this algorithm, two 24models (usually, one high-fidelity and one low-fidelity) are provided to 25Dakota, each with their own \c responses section of the input script, and each 26of which is allowed its own \c simulation_variance. 27The variance specified in the \c responses 28block belonging to the high-fidelity model is applied to any <i> new </i> 29high-fidelity data that is produced by Dakota running the high-fidelity model. 30In the \c experimental_design algorithm, 31low-fidelity model responses are used during the calibration of the model 32parameters, the calculation of the mutual information, and the 33calculation of any posterior statistics after the algorithm is complete. The 34\c simulation_variance is applied to the low-fidelity model responses that are 35used in the calculation of the mutual information. See the User's and Theory 36Manuals for more information. 37 38Topics:: 39Examples:: 40The example below shows two \c responses blocks, one for the low-fidelity 41model and one for the high-fidelity model. Both contain \c simulation_variance 42commands that will apply to the low- and high-fidelity model responses, 43respectively. 44 45\verbatim 46responses, 47 id_responses = 'low-fidelity' 48 calibration_terms = 1 49 simulation_variance = 0.5 50 51responses, 52 id_responses = 'high-fidelity' 53 calibration_terms = 1 54 calibration_data_file = 'dakota_bayes_expdesign.dat' 55 freeform 56 num_config_variables = 1 57 num_experiments = 1 58 experiment_variance_type = 'none' 59 simulation_variance = 1.2 60\endverbatim 61 62Theory:: 63Faq:: 64See_Also:: 65