1Blurb::Pareto set optimization 2Description:: 3In the pareto set minimization method (\c pareto_set), a series of 4optimization or least squares calibration runs are performed for 5different weightings applied to multiple objective functions. This 6set of optimal solutions defines a "Pareto set," which is useful for 7investigating design trade-offs between competing objectives. The 8code is similar enough to the \c multi_start technique that both 9algorithms are implemented in the same ConcurrentMetaIterator class. 10 11The \c pareto_set specification must identify an optimization or least 12squares calibration method using either a \c method_pointer or a \c 13method_name plus optional \c model_pointer. This minimizer is 14responsible for computing a set of optimal solutions from a set of 15response weightings (multi-objective weights or least squares term 16weights). These weightings can be specified as follows: (1) using \c 17random_weight_sets, in which case weightings are selected randomly 18within [0,1] bounds, (2) using \c weight_sets, in which the weighting 19sets are specified in a list, or (3) using both \c random_weight_sets 20and \c weight_sets, for which the combined set of weights will be 21used. In aggregate, at least one set of weights must be specified. 22The set of optimal solutions is called the "pareto set," which can 23provide valuable design trade-off information when there are competing 24objectives. 25 26<b>Expected HDF5 Output</b> 27 28If Dakota was built with HDF5 support and run with the 29\ref environment-results_output-hdf5 keyword, this method 30writes the best parameters and responses returned by each sub-iterator. 31The weights are provided as metadata. See the \ref hdf5_results-ms_pareto 32documentation for details. 33 34 35Topics:: 36Examples:: 37Theory:: 38Faq:: 39See_Also:: 40