Design and analysis of computer experiments

By L. Etman. 2005

How researchers see this paper

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

Citations in the corpus

Taking the Human Out of the Loop: A Review of Bayesian Optimization (2016, background): “[128] (and more recently a book by Santner et al.”
Kernels for Vector-Valued Functions: a Review (2011, background): “The particular input values employed are obtained from different ranges that the input variables can take (given by an expert), and are chosen according to some space-filling criteria (Latin hypercube design, for example) [84].”
ParamILS: An Automatic Algorithm Configuration Framework (2014, methodology): “Sequential parameter optimization (SPO) (Bartz-Beielstein, 2006) is a model-based parameter optimization approach based on the Design and Analysis of Computer Experiments (DACE; see, e.g., Santner, Williams & Notz, 2003), a prominent approach in statistics for blackbox function optimization.”
Probabilistic sensitivity analysis of complex models: a Bayesian approach (2004, background): “The choice of design points is discussed in Sacks et al. (1989).”
Managing computational complexity using surrogate models: a critical review (2020, background): “Many people have used kriging instead of complicated mathematical models, such as (Sacks et al. 1989).”
Efficient Emulators for Multivariate Deterministic Functions (2008, background): “From the early papers of Sacks et al. (1989) and Currin et al. (1991), the technology of emulation has rapidly matured, to the point where there is a standard approach for scalar outputs using a Gaussian process (see, e.g., Santner et al. 2003).”
Combining Field Data and Computer Simulations for Calibration and Prediction (2005, methodology): “A standard prior model for an unknown function is a Gaussian process (GP) (O’Hagan, 1978; Sacks et al., 1989).”
Cases for the nugget in modeling computer experiments (2010, methodology): “The canonical choice of surrogate model for computer experiments is the stationary Gaussian process (Sacks et al. 1989; O’Hagan et al. 1999; Santner et al. 2003), which defines a random process whose evaluation at any finite collection of locations has a multivariate Gaussian distribution with a spe”
A Nonstationary Space-Time Gaussian Process Model for Partially Converged Simulations (2013, methodology): “To assess the global uncertainty of each model, the average prediction variance at test points (referred to as integrated mean square error [IMSE], which is the classical terminology in computer experiments [Sacks et al. (1989)]) and the maximum prediction variance (maxMSE) are computed.”
Kriging metamodeling in simulation: A review (2009, methodology): “, [44]’s computer program in C running under Unix or Linux—called ‘PErK’ (an acronym for ‘Parametric Empirical Kriging’)—allows for white noise, whereas [35]’s ‘DACE’ (the acronym for ‘Design and Analysis of Computer Experiments’) is written in Matlab and is very well documented J.”
Fixed-domain asymptotics for a subclass of Matern-type Gaussian random fields (2005, unknown): “Stein [11] observed that the Gaussian model given by (1) may have some undesirable properties.”
A comprehensive evaluation of various sensitivity analysis methods: A case study with a hydrological model (2014, background): “The most concerned problems in designing a deterministic experiment are whether the sample points are " space-filling " in the design space (Sacks et al., 1989) and how many sample points are sufficient for the experiment (Loeppky et al., 2009).”
Convergence properties of the expected improvement algorithm with fixed mean and covariance functions (2007, unknown): “In the ﬁeld of design and analysis of computer experiments, as pioneered by Sacks et al. (1989) and Currin et al. (1991), the objective function—typically an expensive-to-evaluate numerical model of some physical phenomenon—is seen as a sample path of a stochastic process.”
Automated configuration of algorithms for solving hard computational problems (2009, background): “For example, they can be used to determine how much algorithm response depends on the setting of single parameters, and to which degree parameters interact (Jones et al., 1998; Santner et al., 2003; Bartz-Beielstein, 2006).”