Search results

Large sparse least-squares problems arise in different scientific disciplines such as optimization, data analysis, machine learning and simulation. This paper aims to propose a two-level hybrid direct-iterative scheme, based on novel block independent column reordering, for efficiently solving large sparse least-squares linear systems.

Herewith, a novel block column independent set reordering scheme is used to separate the columns in two groups: columns that are block independent and columns that are coupled. The permutation scheme leads to a two-level hierarchy. Using this two-level hierarchy, the solution of the least-squares linear system results in the solution of a reduced size Schur complement-type square linear system, using the preconditioned conjugate gradient (PCG) method as well as backward substitution using the upper triangular factor, computed through sparse Q-less QR factorization of the columns that are block independent. To improve the convergence behavior of the PCG method, the upper triangular factor, computed through sparse Q-less QR factorization of the coupled columns, is used as a preconditioner. Moreover, to further reduce the fill-in, then the column approximate minimum degree (COLAMD) algorithm is used to permute the block consisting of the coupled columns.

The memory requirements for solving large sparse least-squares linear systems are significantly reduced compared to Q-less QR decomposition of the original as well as the permuted problem with COLAMD. The memory requirements are reduced further by choosing to form larger blocks of independent columns. The convergence behavior of the iterative scheme is improved due to the chosen preconditioning scheme. The proposed scheme is inherently parallel due to the introduction of block independent column reordering.

The proposed scheme is a hybrid direct-iterative approach for solving sparse least squares linear systems based on the implicit computation of a two-level approximate pseudo-inverse matrix. Numerical results indicating the applicability and effectiveness of the proposed scheme are given.

The purpose of this paper is to propose a robust control scheme for near space vehicle's (NSV's) reentry attitude tracking problem under aerodynamic parameter variations and external disturbances.

The robust control scheme is composed of dynamic surface control (DSC) and least squares support vector machines (LS‐SVM). DSC is used to design a nonlinear controller for HSV; then, to increase the robustness and improve the control performance of the controller. LS‐SVM is presented to estimate the lumped uncertainties, including aerodynamic parameter variations and external disturbances. The stability analysis shows that all closed‐loop signals are bounded, with output tracking error and estimate error of LS‐SVM weights exponentially converging to small compacts.

Simulation results demonstrate that the proposed method is effective, leading to promising performance.

First, a robust control scheme composed of DSC and adaptive LS‐SVM is proposed for NSV's reentry attitude tracking problem under aerodynamic parameter variations and external disturbances; second, the proposed method can achieve more favorable tracking performances than conventional dynamic surface control because of employing LS‐SVM to estimate aerodynamic parameter variations and external disturbances.

We investigate the application of a least squares finite element method for the solution of fluid flow problems. The least squares finite element method is based on the minimization of the L2 norm of the equation residuals. Upon discretization, the formulation results in a symmetric, positive definite matrix system which enables efficient iterative solvers to be used. The other motivations behind the development of least squares finite element methods are the applicability of higher order elements and the possibility of using the norm associated to the least squares functional for error estimation. For steady incompressible flows, we develop a method employing linear and quadratic triangular elements and compare their respective accuracy. For steady compressible flows, an implicit conservative least squares scheme which can capture shocks without the addition of artificial viscosity is proposed. A refinement strategy based upon the use of the least squares residuals is developed and several numerical examples are used to illustrate the capabilities of the method when implemented on unstructured triangular meshes.

The purpose of this paper is to probe the recursive identification of piecewise affine Hammerstein models directly by using input-output data. To explain the identification process of a parametric piecewise affine nonlinear function, the authors prove that the inverse function corresponding to the given piecewise affine nonlinear function is also an equivalent piecewise affine form. Based on this equivalent property, during the detailed identification process with respect to piecewise affine function and linear dynamical system, three recursive least squares methods are proposed to identify those unknown parameters under the probabilistic description or bounded property of noise.

First, the basic recursive least squares method is used to identify those unknown parameters under the probabilistic description of noise. Second, multi-innovation recursive least squares method is proposed to improve the efficiency lacked in basic recursive least squares method. Third, to relax the strict probabilistic description on noise, the authors provide a projection algorithm with a dead zone in the presence of bounded noise and analyze its two properties.

Based on complex mathematical derivation, the inverse function of a given piecewise affine nonlinear function is also an equivalent piecewise affine form. As the least squares method is suited under one condition that the considered noise may be a zero mean random signal, a projection algorithm with a dead zone in the presence of bounded noise can enhance the robustness in the parameter update equation.

To the best knowledge of the authors, this is the first attempt at identifying piecewise affine Hammerstein models, which combine a piecewise affine function and a linear dynamical system. In the presence of bounded noise, the modified recursive least squares methods are efficient in identifying two kinds of unknown parameters, so that the common set membership method can be replaced by the proposed methods.

The purpose of this study is to build a high accuracy thrust model under various small turbojet engine shaft speeds by using robust, ordinary, linear and nonlinear least squares estimation methods for target drone applications.

The dynamic shaft speeds from the test experiment of a target drone engine is conducted. Then, thrust values are calculated. Based on these, the engine thrust is modeled by robust linear and nonlinear equations. The models are benefited from quadratic, power and various series expansion functions with several coefficients to optimize the model parameters.

The error terms and accuracy of the model are given using sum of squared errors, root mean square error (RMSE) and R-squared (R²) error definitions. Based on the multiple analyses, it is seen that the RMSE values are no more than 17.7539 and the best obtained result for robust least squares estimation is 15.0086 for linear at all cases. Furthermore, the R² value is found to be 0.9996 as the highest with the nonlinear Fourier series expansion model.

The motivation behind this paper is to propose robust nonlinear thrust models based on power, Fourier and various series expansion functions for dynamic shaft speeds from the test experiments.

The different types of estimators of rational expectations models are surveyed. A key feature is that the model′s solution has to be taken into account when it is estimated. The two ways of doing this, the substitution and errors‐in‐variables methods, give rise to different estimators. In the former case, a generalised least‐squares or maximum‐likelihood type estimator generally gives consistent and efficient estimates. In the latter case, a generalised instrumental variable (GIV) type estimator is needed. Because the substitution method involves more complicated restrictions and because it resolves the solution indeterminacy in a more arbitary fashion, when there are forward‐looking expectations, the errors‐in‐variables solution with the GIV estimator is the recommended combination.

Most studies of power‐law fluids are carried out using stress‐based system of Navier‐Stokes equations; and least‐squares finite element models for vorticity‐based equations of power‐law fluids have not been explored yet. Also, there has been no study of the weak‐form Galerkin formulation using the reduced integration penalty method (RIP) for power‐law fluids. Based on these observations, the purpose of this paper is to fulfill the two‐fold objective of formulating the least‐squares finite element model for power‐law fluids, and the weak‐form RIP Galerkin model of power‐law fluids, and compare it with the least‐squares finite element model.

For least‐squares finite element model, the original governing partial differential equations are transformed into an equivalent first‐order system by introducing additional independent variables, and then formulating the least‐squares model based on the lower‐order system. For RIP Galerkin model, the penalty function method is used to reformulate the original problem as a variational problem subjected to a constraint that is satisfied in a least‐squares (i.e. approximate) sense. The advantage of the constrained problem is that the pressure variable does not appear in the formulation.

The non‐Newtonian fluids require higher‐order polynomial approximation functions and higher‐order Gaussian quadrature compared to Newtonian fluids. There is some tangible effect of linearization before and after minimization on the accuracy of the solution, which is more pronounced for lower power‐law indices compared to higher power‐law indices. The case of linearization before minimization converges at a faster rate compared to the case of linearization after minimization. There is slight locking that causes the matrices to be ill‐conditioned especially for lower values of power‐law indices. Also, the results obtained with RIP penalty model are equally good at higher values of penalty parameters.

Vorticity‐based least‐squares finite element models are developed for power‐law fluids and effects of linearizations are explored. Also, the weak‐form RIP Galerkin model is developed.

The purpose of this paper is to present extended least squares support vector machines (LS‐SVM) where data selection methods are used to get sparse LS‐SVM solution, and to overview and compare the most important data selection approaches.

The selection methods are compared based on their theoretical background and using extensive simulations.

The paper shows that partial reduction is an efficient way of getting a reduced complexity sparse LS‐SVM solution, while partial reduction exploits full knowledge contained in the whole training data set. It also shows that the reduction technique based on reduced row echelon form (RREF) of the kernel matrix is superior when compared to other data selection approaches.

Data selection for getting a sparse LS‐SVM solution can be done in the different representations of the training data: in the input space, in the intermediate feature space, and in the kernel space. Selection in the kernel space can be obtained by finding an approximate basis of the kernel matrix.

The RREF‐based method is a data selection approach with a favorable property: there is a trade‐off tolerance parameter that can be used for balancing complexity and accuracy.

The paper gives contributions to the construction of high‐performance and moderate complexity LS‐SVMs.

Gives introductory remarks about chapter 1 of this group of 31 papers, from ISEF 1999 Proceedings, in the methodologies for field analysis, in the electromagnetic community. Observes that computer package implementation theory contributes to clarification. Discusses the areas covered by some of the papers ‐ such as artificial intelligence using fuzzy logic. Includes applications such as permanent magnets and looks at eddy current problems. States the finite element method is currently the most popular method used for field computation. Closes by pointing out the amalgam of topics.