Surrogate Modeling of Computer Experiments With Different Mesh Densities

Model-based methods for continuous and discrete global optimization

“…Typical discrete data structures include, e.g., ordinal integers, categorical variables, binary variables, permutations, strings, trees or graphs in general. In most of these discrete cases, bounds 50 a and b are also not required anymore.…”

“…Tuo et al [50] proposed a finite-element analysis with its mesh density as the 255 tuning parameter. A problem in casting simulation is used to illustrate this approach.…”

Assessing fire safety using complex numerical models with a Bayesian multi-fidelity approach

“…In this article, we work with a continuum of levels of fidelity indexed by a positive scalar t, following Picheny and Ginsbourger [9] and Tuo et al [14]. The level t can be seen as an input of the simulator, as the input vector x (see Figure 1).…”

“…The prior distribution of the mean response ξ was proposed by Tuo et al [14] for deterministic simulators. To build the prior model, the mean response ξ is assumed to be the sum of an ideal response ξ 0 and a numerical error ε:…”

“…Thus, we assume that ε (x,t = 0) = 0 for all x. We suggest taking a covariance function proposed by Tuo et al [14]:…”