Several articles discuss the design of sample schemes to improve the quality of Kriging estimation. When Kriging is used as part of an iterative response-surface-based global optimization method, the sample scheme is determined by an initial sampling, and the successive evaluation points proposed by the optimization algorithm. The aim of an iterative global optimization algorithm is to construct a sequence converging to a global optimum of the objective function. The larger the share of a subsequence in the tail of the evaluation points, converging to a global optimum, the more the Kriging estimations will be biased. The variance is biased and the estimation mean is biased towards the global optimum.
Original languageEnglish
Title of host publicationAbstract book of the 22nd national conference of the Belgian Operations Research Society - ORBEL 22
Pages10-11
Number of pages2
Publication statusPublished - Jan 2008

Publication series

NameAbstract book of the 22nd national conference of the Belgian Operations Research Society - ORBEL 22

    Research areas

  • Kriging, global optimization, response surface, estimation, sampling

ID: 1784691