On applying Kriging-based approximate optimization to inaccurate data

This paper describes a new approach for the use of Kriging-based approximate optimization with inaccurate data. A conventional Kriging system is applicable to several kinds of complex systems, however, there are some difficulties in applying a Kriging-based system to a noisy system.Therefore, this paper proposes a new approach for Kriging-based estimation with inaccurate data. The principal idea is summarized as an improvement of a semivariogram and determining process of its parameters.By applying the proposed method to some numerical examples, the applicability of the Kriging system to an approximate optimization system of inaccurate data is investigated. The applicability and problems in applying the Kriging system to inaccurate data are discussed, and numerical results illustrate the validity and effectiveness of the proposed method.