Mathematical Problems in Engineering / 2016 / Article / Tab 2

Research Article

An Improved Approach for Estimating the Hyperparameters of the Kriging Model for High-Dimensional Problems through the Partial Least Squares Method

Table 2

Results of the Griewank function in 20D over the interval . Ten trials are done for each test (50, 100, 200, and 300 training points). Best results of the relative error are highlighted in bold type for each case.

SurrogateStatistic50 points100 points
Error (%)CPU timeError (%)CPU time

Kriging Mean0.6230.43 s0.4340.09 s
std0.039.03 s0.0411.96 s
KPLS1Mean0.540.05 s0.530.12 s
std0.030.007 s0.030.02 s
KPLS2 Mean0.520.11 s0.481.04 s
std0.030.05 s0.040.97 s
KPLSMean0.511.27 s0.463.09 s
std0.031.29 s0.063.93 s
KPLS+KMean0.591.20 s0.452.42 s
std0.040.16 s0.070.44 s
KPLS+K Mean0.581.28 s0.423.38 s
std0.040.15 s0.051.06 s
KPLS+K Mean0.582.45 s0.415.61 s
std0.031.32 s0.053.99 s

SurrogateStatistic200 points300 points
Error (%)CPU timeError (%)CPU time

Kriging Mean0.15120.74 s0.1694.31 s
std0.0227.49 s0.0621.92 s
KPLSMean0.480.43 s0.450.89 s
std0.030.08 s0.030.02 s
KPLSMean0.421.14 s0.382.45 s
std0.040.92 s0.041 s
KPLSMean0.373.56 s0.353.52 s
std0.032.75 s0.061.38 s
KPLS+KMean0.208.00 s0.1719.07 s
std0.041.51 s0.073.19 s
KPLS+K Mean0.189.71 s0.1619.89 s
std0.021.29 s0.052.67 s
KPLS+K Mean0.1611.67 s0.1620.49 s
std0.023.88 s0.053.46 s

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.