Abstract:Kriging or Gaussian process models are popular metamodels (surrogate models or emulators) of simulation models; these metamodels give predictors for input combinations that are not simulated. To validate these metamodels for computationally expensive simulation models, the analysts often apply computationally efficient cross-validation. In this paper, we derive new statistical tests for so-called leave-one-out cross-validation. Graphically, we present these tests as scatterplots augmented with confidence inter… Show more
“…There exist several methods for constructing a confidence bound for the unknown mean response function, including the classical pointwise confidence interval [34], the simultaneous confidence region relying on bootstrapping or the Bonferroni [24] and Šidák corrections [13], and the uniform confidence bounds derived either using the frequentist kernel methods [23] or from the Bayesian GP modeling perspective [43]. We adopt the uniform bound for heteroscedastic metamodeling approaches (including SK) proposed by Kirschner and Krause [23] which holds true with a prescribed high probability across the input space and through all iterations .…”
This paper proposes two sequential metamodel‐based methods for level‐set estimation (LSE) that leverage the uniform bound built on stochastic kriging: predictive variance reduction (PVR) and expected classification improvement (ECI). We show that PVR and ECI possess desirable theoretical performance guarantees and provide closed‐form expressions for their respective sequential sampling criteria to seek the next design point for performing simulation runs, allowing computationally efficient one‐iteration look‐ahead updates. To enhance understanding, we reveal the connection between PVR and ECI's sequential sampling criteria. Additionally, we propose integrating a budget allocation feature with PVR and ECI, which improves computational efficiency and potentially enhances robustness to the impacts of heteroscedasticity. Numerical studies demonstrate the superior performance of the proposed methods compared to state‐of‐the‐art benchmarking approaches when given a fixed simulation budget, highlighting their effectiveness in addressing LSE problems.
“…There exist several methods for constructing a confidence bound for the unknown mean response function, including the classical pointwise confidence interval [34], the simultaneous confidence region relying on bootstrapping or the Bonferroni [24] and Šidák corrections [13], and the uniform confidence bounds derived either using the frequentist kernel methods [23] or from the Bayesian GP modeling perspective [43]. We adopt the uniform bound for heteroscedastic metamodeling approaches (including SK) proposed by Kirschner and Krause [23] which holds true with a prescribed high probability across the input space and through all iterations .…”
This paper proposes two sequential metamodel‐based methods for level‐set estimation (LSE) that leverage the uniform bound built on stochastic kriging: predictive variance reduction (PVR) and expected classification improvement (ECI). We show that PVR and ECI possess desirable theoretical performance guarantees and provide closed‐form expressions for their respective sequential sampling criteria to seek the next design point for performing simulation runs, allowing computationally efficient one‐iteration look‐ahead updates. To enhance understanding, we reveal the connection between PVR and ECI's sequential sampling criteria. Additionally, we propose integrating a budget allocation feature with PVR and ECI, which improves computational efficiency and potentially enhances robustness to the impacts of heteroscedasticity. Numerical studies demonstrate the superior performance of the proposed methods compared to state‐of‐the‐art benchmarking approaches when given a fixed simulation budget, highlighting their effectiveness in addressing LSE problems.
“…To get the distribution of emergency rescue response level in the study area, the normalized data need to be quantified spatially. This study uses the Kriging interpolation method for spatial quantization (Kleijnen and van Beers, 2022;Meng, 2021). This interpolation method has the following two advantages:…”
Abstract. Earthquake disasters often cause massive casualties and property losses, so it is essential to achieve rapid emergency rescue response. Although many scholars have studied the speed of emergency rescue response from different perspectives, there are still many problems in the impact of road congestion on travel paths and distances, as well as the timeliness of the road network. Therefore, this study calculated the travel distance from the rescue point to the disaster point using the path planning method based on the Baidu map navigation data, and obtained the distribution map of the emergency rescue response level in Wenchuan County, Sichuan Province, China. The results show that :(1) the path planning method based on Baidu map navigation data is more realistic and accurate than the traditional method. (2) there are apparent spatial differences in the emergency rescue response in Wenchuan County, and the emergency rescue response level in Weizhou town in the north is higher than that in Wolong and Sanjiang town in the southwest. The research results have significant reference value for planning emergency rescue routes and improving rescue speed in Wenchuan County and similar counties.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.