The option has an effectįor types fact and faure only. Logical that prevents randomization per default. If all factors have the same scaling that deviates from the default 0/1 scale ends. Interpreted as the number of levels for type fact, if nlevels is Of levels, the number of levels can also be specified through nruns, which is Integers specifies the number of levels for each factor. Used for type fact only integer number or numeric vector of nfactor the unit cube generated by the functions from packages lhs Note that the rounding is applied after generation of the design on the actualĭata scale, i.e. (the same for all factors) or a vector of length nfactors Is rescaled to the values given in factor.names.ĭigits to which the design columns are rounded one single value The original unit cube calculated by package lhs (scale ends 0 and 1 for each variable) If the list is not named, the variable names are X1, X2 and so forth Of standard second order analysis methods on the resulting data ( link) List of scale end values for each factor Results are reproducible within a package version, but reproducibility between Specifying a seed used to make the result reproducible for Seed for random number generation latin hypercube samples from package See the respective functions from packages lhs and Where possible choices for type are augment, optSeeded, or optAugment. Or function runif.faure from package DiceDesign (type faure).įunction lhs.augment calls function typeLHS from package lhs, Lhs (types genetic, improved, maximin, optimum, random),Ī function named typeDesign from package DiceDesign (types dmax, strauss, fact) Number of factors in the latin hypercube sampleĬharacter string indicating the type of design or augmentation method ĭefaults are “optimum” for lhs.design and “optAugment” Nruns can be missing or can be the correctly-specified number of runs. the resulting design will have nruns^nfactors runs Īlternatively, if nlevels is separately specified as a vector of different numbers of levels, This number is taken to be the common number of levels of all factors, Number of runs in the latin hypercube sample įor type fact (a full factorial with equally-space levels), ) lhs.augment ( lhs, m = 1, type = "optAugment", seed = NULL. Through comparative studies on sampling property and metamodel accuracy, the new approach is shown to outperform other sequential sampling methods for global metamodelling and is comparable to the one-stage sampling method while providing more flexibility in a sequential metamodelling ( nruns, nfactors, type = "optimum", factor.names = NULL, seed = NULL, digits = NULL, nlevels = nruns, default.levels = c ( 0, 1 ), randomize = FALSE. The sequential sampling is formulated as an optimization problem, with the objective being the Maximin Distance, a space-filling criterion, and the constraints based on a set of pre-specified minimum one-dimensional distances to achieve the approximate one-dimensional projective property. The goal in this article is to develop an efficient and effective sequential Quasi-LHD (Latin Hypercube design) sampling method to maintain and balance the two aforementioned properties. Through comparative studies on sampling property and metamodel accuracy, the new approach is shown to outperform other sequential sampling methods for global metamodelling and is comparable to the one-stage sampling method while providing more flexibility in a sequential metamodelling procedure.ĪB - Space-filling and projective properties are desired features in the design of computer experiments to create global metamodels to replace expensive computer simulations in engineering design. N2 - Space-filling and projective properties are desired features in the design of computer experiments to create global metamodels to replace expensive computer simulations in engineering design. The views expressed are those of the authors and do not necessarily reflect the views of the sponsors. The grant support from National Science Foundation (CMMI – 0522662) and the China Scholarship Council are greatly acknowledged. T1 - Optimizing latin hypercube design for sequential sampling of computer experiments
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |