1Blurb:: 2Select a search method for Newton-based optimizers 3Description:: 4The \c search_method control is defined for all Newton-based 5optimizers and is used to select between \c trust_region, \c 6gradient_based_line_search, and \c value_based_line_search methods. 7The \c gradient_based_line_search option uses the line search method 8proposed by \cite More1994. This option 9satisfies sufficient decrease and curvature conditions; whereas, \c 10value_base_line_search only satisfies the sufficient decrease 11condition. At each line search iteration, the \c 12gradient_based_line_search method computes the function and gradient 13at the trial point. Consequently, given expensive function 14evaluations, the \c value_based_line_search method is preferred to the 15\c gradient_based_line_search method. Each of these Newton methods 16additionally supports the \c tr_pds selection for unconstrained 17problems. This option performs a robust trust region search using 18pattern search techniques. Use of a line search is the default for 19bound-constrained and generally-constrained problems, and use of a \c 20trust_region search method is the default for unconstrained problems. 21 22 23Topics:: 24Examples:: 25Theory:: 26Faq:: 27See_Also:: 28