A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output From a Computer Code

By M. D. McKay, R. Beckman, William J. Conover. 2000

12,710 citations recorded by the source corpus.

Subfields: Finance, Management Science and Operations Research, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Finance

In Management Science and Operations Research

In Statistics and Probability

Citations in the corpus

Bayesian Neural Architecture Search using A Training-Free Performance Metric (2020, methodology): “After creating the initial data set of BO ( X, Y ) using Latin Hypercube Sampling [31], the user can choose to turn on the generation of the warm-data prior to the optimization loop (lines 6-9).”
An evaluation of adaptive surrogate modeling based optimization with two benchmark problems (2014, methodology): “This method was first described by McKay et al. (1979) and further elaborated by Iman et al. (1981) and Hernandez et al. (2012).”
Physics informed deep learning for computational elastodynamics without labeled data (2020, background): “Note that the collocation points can be generated via the Latin hypercube sampling (LHS) strategy [53].”
Multi-fidelity optimization via surrogate modelling (2007, methodology): “The definition of ‘uniform coverage’ is by no means obvious and a substantial body of literature exists on the subject—here we work with the optimality criterion of Morris & Mitchell (1995). According to this, the plan with the best space-filling properties is that which maximizes the smallest dista”
Combining Field Data and Computer Simulations for Calibration and Prediction (2005, background): “experimental design – determining input settings at which to carry out a sequence of simulation designs ( McKay et al., 1979; Johnson et al., 1990; et al., Mitchell et al., 1995);”
Random Search for Hyper-Parameter Optimization (2012, background): “…the Sobol (Antonov and Saleev, 1979), Halton (Halton, 1960), and Niederreiter (Bratley et al., 1992) sequences, as well as latin hypercube sampling (McKay et al., 1979) are all more or less deterministic schemes for getting point sets that are more representative of random uniform draws than…”
Sensitivity and uncertainty analysis coupled with automatic calibration for a distributed watershed model (2005, background): “However, the use of more strategic, efficient, and effective sampling approaches, such as importance sampling and Latin hypercube sampling, can significantly reduce computational demand (McKay et al., 1979; Iman and Conover, 1980).”
Borg: An Auto-Adaptive Many-Objective Evolutionary Computing Framework (2013, methodology): “Each algorithm was executed 1,000 times using parameters produced by a Latin hypercube sampling (LHS; McKay et al., 1979) across each algorithm’s feasible parameter range.”
Calculations of Sobol indices for the Gaussian process metamodel (2008, background): “In practical situations, a metamodel with a predictivity lower than 0 .”
Efficient Pareto Frontier Exploration using Surrogate Approximations (2000, background): “A variety of different types of experimental designs exist, including central composite designs (Myers and Montgomery, 1995), Latin hypercubes (McKay et al., 1979), and orthogonal arrays (Owen, 1992) to name a few.”
Critical Hyper-Parameters: No Random, No Cry (2017, background): “LHS [6, 16] construct a sequence as follows: for a given n , consider the partition of [0 , 1] d into a regular grid of n d cells and then deﬁne the index of a cell as its position in the corresponding n d grid; choose d random permutations σ 1 , . . . , σ d of { 1 , . . . , n } and choose for each…”