Supersaturated design is essentially a fractional factorial design in which the number of potential effects is greater than the number of runs. In this article, the supersaturated design is applied to a computer exper...Supersaturated design is essentially a fractional factorial design in which the number of potential effects is greater than the number of runs. In this article, the supersaturated design is applied to a computer experiment through an example of steady current circuit model problem. A uniform mixed-level supersaturated design and the centered quadratic regression model are used. This example shows that supersaturated design and quadratic regression modeling method are very effective for screening effects and building the predictor. They are not only useful in computer experiments but also in industrial and other scientific experiments.展开更多
Latin hypercube design and uniform design are two kinds of most popular space-filling designs for computer experiments. The fact that the run size equals the number of factor levels in a Latin hypercube design makes i...Latin hypercube design and uniform design are two kinds of most popular space-filling designs for computer experiments. The fact that the run size equals the number of factor levels in a Latin hypercube design makes it difficult to be orthogonal. While for a uniform design, it usually has good space-filling properties, but does not necessarily have small or zero correlations between factors. In this paper, we construct a class of column-orthogonal and nearly column-orthogonal designs for computer experiments by rotating groups of factors of orthogonal arrays, which supplement the designs for computer experiments in terms of various run sizes and numbers of factor levels and are flexible in accommodating various combinations of factors with different numbers of levels. The resulting column-orthogonal designs not only have uniformly spaced levels for each factor but also have uncorrelated estimates of the linear effects in first order models. Further, they are 3-orthogonal if the corresponding orthogonal arrays have strength equal to or greater than three. Along with a large factor-to-run ratio, these newly constructed designs are economical and suitable for screening factors for physical experiments.展开更多
In this paper,we propose a new method,called the level-collapsing method,to construct branching Latin hypercube designs(BLHDs).The obtained design has a sliced structure in the third part,that is,the part for the shar...In this paper,we propose a new method,called the level-collapsing method,to construct branching Latin hypercube designs(BLHDs).The obtained design has a sliced structure in the third part,that is,the part for the shared factors,which is desirable for the qualitative branching factors.The construction method is easy to implement,and(near)orthogonality can be achieved in the obtained BLHDs.A simulation example is provided to illustrate the effectiveness of the new designs.展开更多
Strong orthogonal arrays(SOAs) were recently introduced and studied as a class of spacefilling designs for computer experiments. To surely realize better space-filling properties, SOAs of strength three or higher are ...Strong orthogonal arrays(SOAs) were recently introduced and studied as a class of spacefilling designs for computer experiments. To surely realize better space-filling properties, SOAs of strength three or higher are desirable. In addition, orthogonality is also an important property for designs of computer experiments, because it guarantees that the estimates of the main effects are uncorrelated. This paper first provides a systematic study on the construction of(nearly) orthogonal strength-three SOAs with better space-filling properties. The newly proposed strength-three SOAs enjoy almost the same space-filling properties of strength-four SOAs, and can accommodate much more columns than the latter. Moreover, they are(nearly) orthogonal and flexible in run sizes. The construction methods are straightforward to implement, and their theoretical supports are well established. In addition to the theoretical results, many designs are tabulated for practical needs.展开更多
Space-filling designs are widely used in computer experiments.They are frequently evaluated by the orthogonality and distance-related criteria.Rotating orthogonal arrays is an appealing approach to constructing orthog...Space-filling designs are widely used in computer experiments.They are frequently evaluated by the orthogonality and distance-related criteria.Rotating orthogonal arrays is an appealing approach to constructing orthogonal space-filling designs.An important issue that has been rarely addressed in the literature is the design selection for the initial orthogonal arrays.This paper studies the maximin L_(2)-distance properties of orthogonal designs generated by rotating two-level orthogonal arrays under three criteria.We provide theoretical justifications for the rotation method from a maximin distance perspective and further propose to select initial orthogonal arrays by the minimum G_(2)-aberration criterion.New infinite families of orthogonal or 3-orthogonal U-type designs,which also perform well under the maximin distance criterion,are obtained and tabulated.Examples are presented to show the effectiveness of the constructed designs for building statistical surrogate models.展开更多
Latin hypercube designs(LHDs)are very popular in designing computer experiments.In addition,orthogonality is a desirable property for LHDs,as it allows the estimates of the main effects in linear models to be uncorrel...Latin hypercube designs(LHDs)are very popular in designing computer experiments.In addition,orthogonality is a desirable property for LHDs,as it allows the estimates of the main effects in linear models to be uncorrelated with each other,and is a stepping stone to the space-filling property for fitting Gaussian process models.Among the available methods for constructing orthogonal Latin hypercube designs(OLHDs),the rotation method is particularly attractive due to its theoretical elegance as well as its contribution to spacefilling properties in low-dimensional projections.This paper proposes a new rotation method for constructing OLHDs and nearly OLHDs with flexible run sizes that cannot be obtained by existing methods.Furthermore,the resulting OLHDs are improved in terms of the maximin distance criterion and the alias matrices and a new kind of orthogonal designs are constructed.Theoretical properties as well as construction algorithms are provided.展开更多
Quasi-regression, motivated by the problems arising in the computer experiments, focuses mainly on speeding up evaluation. However, its theoretical properties are unexplored systemically. This paper shows that quasi-r...Quasi-regression, motivated by the problems arising in the computer experiments, focuses mainly on speeding up evaluation. However, its theoretical properties are unexplored systemically. This paper shows that quasi-regression is unbiased, strong convergent and asymptotic normal for parameter estimations but it is biased for the fitting of curve. Furthermore, a new method called unbiased quasi-regression is proposed. In addition to retaining the above asymptotic behaviors of parameter estimations, unbiased quasi-regression is unbiased for the fitting of curve.展开更多
基金Research supported by the National Natural Science Foundation of China (10301015)the Science and Technology Innovation Fund of Nankai University, the Visiting Scholar Program at Chern Institute of Mathematicsa Hong Kong Research Grants Council Grant (RGC/HKBU 200804)
文摘Supersaturated design is essentially a fractional factorial design in which the number of potential effects is greater than the number of runs. In this article, the supersaturated design is applied to a computer experiment through an example of steady current circuit model problem. A uniform mixed-level supersaturated design and the centered quadratic regression model are used. This example shows that supersaturated design and quadratic regression modeling method are very effective for screening effects and building the predictor. They are not only useful in computer experiments but also in industrial and other scientific experiments.
基金supported by the Program for New Century Excellent Talents in Universityof China (Grant No. NCET-07-0454)National Natural Science Foundation of China (Grant No. 10971107)the Fundamental Research Funds for the Central Universities (Grant No. 10QNJJ003)
文摘Latin hypercube design and uniform design are two kinds of most popular space-filling designs for computer experiments. The fact that the run size equals the number of factor levels in a Latin hypercube design makes it difficult to be orthogonal. While for a uniform design, it usually has good space-filling properties, but does not necessarily have small or zero correlations between factors. In this paper, we construct a class of column-orthogonal and nearly column-orthogonal designs for computer experiments by rotating groups of factors of orthogonal arrays, which supplement the designs for computer experiments in terms of various run sizes and numbers of factor levels and are flexible in accommodating various combinations of factors with different numbers of levels. The resulting column-orthogonal designs not only have uniformly spaced levels for each factor but also have uncorrelated estimates of the linear effects in first order models. Further, they are 3-orthogonal if the corresponding orthogonal arrays have strength equal to or greater than three. Along with a large factor-to-run ratio, these newly constructed designs are economical and suitable for screening factors for physical experiments.
基金supported by the National Natural Science Foundation of China (11601367,11771219 and 11771220)National Ten Thousand Talents Program+1 种基金Tianjin Development Program for Innovation and EntrepreneurshipTianjin "131" Talents Program
文摘In this paper,we propose a new method,called the level-collapsing method,to construct branching Latin hypercube designs(BLHDs).The obtained design has a sliced structure in the third part,that is,the part for the shared factors,which is desirable for the qualitative branching factors.The construction method is easy to implement,and(near)orthogonality can be achieved in the obtained BLHDs.A simulation example is provided to illustrate the effectiveness of the new designs.
基金supported by the National Natural Science Foundation of China under Grant Nos. 12131001and 12226343the MOE Project of Key Research Institute of Humanities and Social Sciences under Grant No.22JJD110001the National Ten Thousand Talents Program of China。
文摘Strong orthogonal arrays(SOAs) were recently introduced and studied as a class of spacefilling designs for computer experiments. To surely realize better space-filling properties, SOAs of strength three or higher are desirable. In addition, orthogonality is also an important property for designs of computer experiments, because it guarantees that the estimates of the main effects are uncorrelated. This paper first provides a systematic study on the construction of(nearly) orthogonal strength-three SOAs with better space-filling properties. The newly proposed strength-three SOAs enjoy almost the same space-filling properties of strength-four SOAs, and can accommodate much more columns than the latter. Moreover, they are(nearly) orthogonal and flexible in run sizes. The construction methods are straightforward to implement, and their theoretical supports are well established. In addition to the theoretical results, many designs are tabulated for practical needs.
基金supported by National Natural Science Foundation of China(Grant Nos.11901199 and 71931004),supported by National Natural Science Foundation of China(Grant Nos.11971098 and 11471069)the Open Research Fund of Key Laboratory for Applied Statistics of Ministry of Education,Northeast Normal University(Grant No.130028906)+1 种基金Shanghai Chenguang Program(Grant No.19CG26)National Key R&D Program of China(Grant No.2020YFA0714102)。
文摘Space-filling designs are widely used in computer experiments.They are frequently evaluated by the orthogonality and distance-related criteria.Rotating orthogonal arrays is an appealing approach to constructing orthogonal space-filling designs.An important issue that has been rarely addressed in the literature is the design selection for the initial orthogonal arrays.This paper studies the maximin L_(2)-distance properties of orthogonal designs generated by rotating two-level orthogonal arrays under three criteria.We provide theoretical justifications for the rotation method from a maximin distance perspective and further propose to select initial orthogonal arrays by the minimum G_(2)-aberration criterion.New infinite families of orthogonal or 3-orthogonal U-type designs,which also perform well under the maximin distance criterion,are obtained and tabulated.Examples are presented to show the effectiveness of the constructed designs for building statistical surrogate models.
基金supported by National Natural Science Foundation of China(Grant Nos.12131001 and 11871288)National Ten Thousand Talents Program and the 111 Project B20016。
文摘Latin hypercube designs(LHDs)are very popular in designing computer experiments.In addition,orthogonality is a desirable property for LHDs,as it allows the estimates of the main effects in linear models to be uncorrelated with each other,and is a stepping stone to the space-filling property for fitting Gaussian process models.Among the available methods for constructing orthogonal Latin hypercube designs(OLHDs),the rotation method is particularly attractive due to its theoretical elegance as well as its contribution to spacefilling properties in low-dimensional projections.This paper proposes a new rotation method for constructing OLHDs and nearly OLHDs with flexible run sizes that cannot be obtained by existing methods.Furthermore,the resulting OLHDs are improved in terms of the maximin distance criterion and the alias matrices and a new kind of orthogonal designs are constructed.Theoretical properties as well as construction algorithms are provided.
基金Project supported by the National Natural Science Foundation of China (No. 10571093, No. 10371059)Specialized Research Fund for the Doctoral Program of Higher Education of China (No. 20050055038)+2 种基金the Natural Science Foundation of Shandong Province of China (No. 2006A13)the China Postdoctoral Science Foundation (No. 20060390169)the Tianjin Planning Programs of Philosophy and Social Science of China (No. TJ05-TJ002).
文摘Quasi-regression, motivated by the problems arising in the computer experiments, focuses mainly on speeding up evaluation. However, its theoretical properties are unexplored systemically. This paper shows that quasi-regression is unbiased, strong convergent and asymptotic normal for parameter estimations but it is biased for the fitting of curve. Furthermore, a new method called unbiased quasi-regression is proposed. In addition to retaining the above asymptotic behaviors of parameter estimations, unbiased quasi-regression is unbiased for the fitting of curve.
