Decomposition and Approximation of Multivariate Functions on the Cube

In this paper, we present a decomposition method of multivariate functions. This method shows that any multivariate function f on [0, 1]d is a finite sum of the form ∑jФjψj, where each Фj can be extended to a smooth periodic function, each ψj is an algebraic polynomial, and each Фjψj is a product of separated variable type and its smoothness is same as f. Since any smooth periodic function can be approximated well by trigonometric polynomials, using our decomposition method, we find that any smooth multivariate function on [0, 1]d can be approximated well by a combination of algebraic polynomials and trigonometric polynomials. Meanwhile, we give a precise estimate of the approximation error.

RELATIVE WIDTH OF SMOOTH CLASSES OF MULTIVARIATE PERIODIC FUNCTIONS WITH RESTRICTIONS ON ITERATED LAPLACE DERIVATIVES IN THE L_2-METRIC

For two subsets W and V of a Banach space X, let Kn（W, V, X） denote the relative Kolmogorov n-width of W relative to V defined by Kn（W,V,X）：=inf sup inf/Ln f∈W g∈V∩Ln‖f-g‖x, where the infimum is taken over all n-dimensional linear subspaces Ln of X. Let W2（△^τ） denote the class of 2π-periodic functions f with d-variables satisfying ∫[-π, π]^d｜△^τf（x）｜^2dx≤1, while △^τ is the r-iterate of Laplace operator △. This article discusses the relative Kolmogorov n-width of W2（△^τ） relative to W2（△^τ） in Lq（[-π, π]^d） （1≤ q ≤ ∞）, and obtain its weak asymptotic result.

The Average Widths and Non-linear Widths of the Classes of Multivariate Functions with Bounded Moduli of Smoothness

The classes of the multivariate functions with bounded moduli on Rd and Td are given and their average σ-widths and non-linear n-widths are discussed. The weak asymptotic behaviors are established for the corresponding quantities.

A Multivariate Empirical Orthogonal Function-Based Scheme for the Balanced Initial Ensemble Generation of an Ensemble Kalman Filter

The initial ensemble perturbations for an ensemble data assimilation system are expected to reasonably sample model uncertainty at the time of analysis to further reduce analysis uncertainty. Therefore, the careful choice of an initial ensemble perturbation method that dynamically cycles ensemble perturbations is required for the optimal performance of the system. Based on the multivariate empirical orthogonal function (MEOF) method, a new ensemble initialization scheme is developed to generate balanced initial perturbations for the ensemble Kalman filter (EnKF) data assimilation, with a reasonable consideration of the physical relationships between different model variables. The scheme is applied in assimilation experiments with a global spectral atmospheric model and with real observations. The proposed perturbation method is compared to the commonly used method of spatially-correlated random perturbations. The comparisons show that the model uncertainties prior to the first analysis time, which are forecasted from the balanced ensemble initial fields, maintain a much more reasonable spread and a more accurate forecast error covariance than those from the randomly perturbed initial fields. The analysis results are further improved by the balanced ensemble initialization scheme due to more accurate background information. Also, a 20-day continuous assimilation experiment shows that the ensemble spreads for each model variable are still retained in reasonable ranges without considering additional perturbations or inflations during the assimilation cycles, while the ensemble spreads from the randomly perturbed initialization scheme decrease and collapse rapidly.

A LAW OF THE ITERATED LOGARITHM FOR NEAREST NEIGHBOR ESTIMATION OF MULTIVARIATE DENSITY FUNCTION

Let X be a d-dimensional random vector with unknown density function f(z) = f (z1, ..., z(d)), and let f(n) be teh nearest neighbor estimator of f proposed by Loftsgaarden and Quesenberry (1965). In this paper, we established the law of the iterated logarithm of f(n) for general case of d greater-than-or-equal-to 1, which gives the exact pointwise strong convergence rate of f(n).

Multivariate Rational Response Surface Approximation of Nodal Displacements of Truss Structures

Polynomial-basis response surface method has some shortcomings for truss structures in structural optimization,concluding the low fitting accuracy and the great computational effort. Based on the theory of approximation, a response surface method based on Multivariate Rational Function basis(MRRSM) is proposed. In order to further reduce the computational workload of MRRSM, focusing on the law between the cross-sectional area and the nodal displacements of truss structure, a conjecture that the determinant of the stiffness matrix and the corresponding elements of adjoint matrix involved in displacement determination are polynomials with the same order as their respective matrices, each term of which is the product of cross-sectional areas, is proposed. The conjecture is proved theoretically for statically determinate truss structure, and is shown corrected by a large number of statically indeterminate truss structures. The theoretical analysis and a large number of numerical examples show that MRRSM has a high fitting accuracy and less computational effort. Efficiency of the structural optimization of truss structures would be enhanced.

WIDTHS AND AVERAGE WIDTHS OF SOBOLEV CLASSES

This paper concerns the problem of the Kolmogorov n-width, the linear re-width, the Gel'fand n-width and the Bernstein re-width of Sobolev classes of the periodic multivariate functions in the space Lp(Td) and the average Bernstein o-width, average Kolmogorov o-widths, the average linear o-widths of Sobolev classes of the multivariate functions in the space LP(R ), where p = (p1,…,pd), 1 < Pj < ∞o, j = 1,2,…,d, or pj = ∞,j = 1,2,…, d. Their weak asymptotic behaviors are established for the corresponding quantities.

FINGERPRINT-BASED KEY BINDING/RECOVERING SCHEME BASED ON FUZZY VAULT

This letter proposes fingerprint-based key binding/recovering with fuzzy vault. Fingerprint minutiae data and the cryptographic key are merged together by a multivariable linear function. First, the minutiae data are bound by a set of random data through the linear function. The number of the function’s variables is determined by the required number of matched minutiae. Then, a new key de- rived from the random data is used to encrypt the cryptographic key. Lastly, the binding data are protected using fuzzy vault scheme. The proposed scheme provides the system with the flexibility to use changeable number of minutiae to bind/recover the protected key and a unified method regardless of the length of the key.

AN IMPROVED MULTIVARIATE LOSS FUNCTION APPROACH TO OPTIMIZATION

The basic purpose of a quality loss function is to evaluate a loss to customers in a quantitativemanner.Although there are several multivariate loss functions that have been proposed and studied inthe literature,it has room for improvement.A good multivariate loss function should represent anappropriate compromise in terms of both process economics and the correlation structure amongvarious responses.More important,it should be easily understood and implemented in practice.According to this criterion,we first introduce a pragmatic dimensionless multivariate loss functionproposed by Artiles-Leon,then we improve the multivariate loss function in two respects:one ismaking it suitable for all three types of quality characteristics;the other is considering correlationstructure among the various responses,which makes the improved multivariate loss function moreadequate in the real world.On the bases of these,an example from industrial practice is provided tocompare our improved method with other methods,and last,some reviews are presented inconclusion.

MULTIVARIATE FOURIER TRANSFORM METHODS OVER SIMPLEX AND SUPER-SIMPLEX DOMAINS

In this paper we propose the well-known Fourier method on some non-tensor product domains in Rd, including simplex and so-called super-simplex which consists of （d ＋ 1）! simplices. As two examples, in 2-D and 3-D case a super-simplex is shown as a parallel hexagon and a parallel quadrilateral dodecahedron, respectively. We have extended most of concepts and results of the traditional Fourier methods on multivariate cases, such as Fourier basis system, Fourier series, discrete Fourier transform （DFT） and its fast algorithm （FFT） on the super-simplex, as well as generalized sine and cosine transforms （DST, DCT） and related fast algorithms over a simplex. The relationship between the basic orthogonal system and eigen-functions of a LaDlacian-like operator over these domains is explored.

L_2-Optimal Recovery on the Riesz Potential Spaces

The problems of best reconstruction of multivariate functions of the Riesz potential spaces from their values on a given mesh are considered, and the exact results of some classes of L_2(R^n) (and L_2(Q^n)) defined by the Riesz potential are obtained.

The essential order of approximation for neural networks

There have been various studies on approximation ability of feedforward neural networks (FNNs). Most of the existing studies are, however, only concerned with density or upper bound estimation on how a multivariate function can be approximated by an FNN, and consequently, the essential approximation ability of an FNN cannot be revealed. In this paper, by establishing both upper and lower bound estimations on approximation order, the essential approximation ability (namely, the essential approximation order) of a class of FNNs is clarified in terms of the modulus of smoothness of functions to be approximated. The involved FNNs can not only approximate any continuous or integrable functions defined on a compact set arbitrarily well, but also provide an explicit lower bound on the number of hidden units required. By making use of multivariate approximation tools, it is shown that when the functions to be approximated are Lipschitzian with order up to 2, the approximation speed of the FNNs is uniquely determined by modulus of smoothness of the functions.

The essential order of approximation for nearly exponential type neural networks

For the nearly exponential type of feedforward neural networks （neFNNs）, it is revealed the essential order of their approximation. It is proven that for any continuous function defined on a compact set of Rd, there exists a three-layer neFNNs with fixed number of hidden neurons that attain the essential order. When the function to be approximated belongs to the α-Lipschitz family （0 〈α≤ 2）, the essential order of approxi- mation is shown to be O（n^-α） where n is any integer not less than the reciprocal of the predetermined approximation error. The upper bound and lower bound estimations on approximation precision of the neFNNs are provided. The obtained results not only characterize the intrinsic property of approximation of the neFNNs, but also uncover the implicit relationship between the precision （speed） and the number of hidden neurons of the neFNNs.

The Irreducibility and Isomorphism of Piecewise Algebraic Sets

In this paper , we define the piecewise algebraic sets by using multivariatespline functions and discuss their irreducibility and isomorphism problem. We presenttwo equivalent conditions for the irreducibility of piecewise algebraic sets, and turn theisomorphism and classifying problem of piecewise algebraic sets into the isonorphismand classifying problem of commutative algebras.

DESIGNING THE OPTIMUM CONFIGURATIONS OF CIRCULAR AND SPHERICAL PRODUCT SPECIFICATIONS FOR MULTIPLE QUALITY CHARACTERISTICS

As an integral part of tolerance design in the context of design for six sigma, determining optimal product specifications has become the focus of increased activity, as manufacturing industries strive to increase productivity and improve the quality of their products. Although a number of research papers have been reported in the research community, there is room for improvement. Most existing research papers consider determining optimal specification limits for a single quality characteristic. In this paper, we develop the modeling and optimization procedures for optimum circular and spherical configurations by considering multiple quality characteristics. The concepts of multivariate quality loss function and truncated distribution are incorporated. This has never been adequately addressed, nor has been appropriately applied in industry. A numerical example is shown and comparison studies are made.

Feasibility study of online tuning of the luminosity in a circular collider with the robust conjugate direction search method

The robust conjugate direction search （RCDS） method has high tolerance to noise in beam experiments. It has been demonstrated that this method can be used to optimize the machine performance of a light source online. In our study, taking BEPCII as an example, the feasibility of online tuning of the luminosity in a circular collider is explored, through numerical simulation and preliminary online experiments. It is shown that the luminosity that is artificially decreased by a deviation of beam orbital offset from optimal trajectory can be recovered with this method.

Numerical Inversion for the Initial Distribution in the Multi-TermTime-FractionalDiffusion Equation Using Final Observations

This article deals with numerical inversion for the initial distribution in the multi-term time-fractional diffusion equation using final observations.The inversion problem is of instability,but it is uniquely solvable based on the solution’s expression for the forward problem and estimation to the multivariate Mittag-Leffler function.From view point of optimality,solving the inversion problem is transformed to minimizing a cost functional,and existence of a minimum is proved by the weakly lower semi-continuity of the functional.Furthermore,the homotopy regularization algorithm is introduced based on the minimization problem to perform numerical inversions,and the inversion solutions with noisy data give good approximations to the exact initial distribution demonstrating the efficiency of the inversion algorithm.

Evaluation of the Madden–Julian Oscillation in Fengyun-3B Polar-Orbiting Satellite Reprocessed OLR Data

The present study compares the spatial and temporal characteristics of the Madden-Julian Oscillation(MJO)in Fengyun-3B(FY-3B)polar-orbiting satellite reprocessed outgoing longwave radiation(OLR)data and NOAA OLR data during 2011-2020.The spatial distributions of climatological mean and intraseasonal standard deviation of FY-3B OLR during boreal winter(November-April)and boreal summer(May-October)are highly consistent with those of NOAA OLR.The FY-3B and NOAA OLRs display highly consistent features in the wavenumber-frequency spectra,the occurrence frequency of MJO active days,the eastward propagation of MJO along the equator,and the interannual variability of MJO according to diagnoses using the all-season multivariate EOF analysis.These results indicate that the FY-3B OLR produced by the polar-orbiting satellites is of high quality and worthy of global application.