Tensor methods for large sparse systems of nonlinear equations

This paper introduces tensor methods for solving large sparse systems of nonlinear equations. Tensor methods for nonlinear equations were developed in the context of solving small to medium-sized dense problems. They base each iteration on a quadratic model of the nonlinear equations, where the second-order term is selected so that the model requires no more derivative or function information per iteration than standard linear model-based methods, and hardly more storage or arithmetic operations per iteration. Computational experiments on small to medium-sized problems have shown tensor methods to be considerably more efficient than standard Newton-based methods, with a particularly large advantage on singular problems. This paper considers the extension of this approach to solve large sparse problems. The key issue considered is how to make efficient use of sparsity in forming and solving the tensor model problem at each iteration. Accomplishing this turns out to require an entirely new way of solving the tensor model that successfully exploits the sparsity of the Jacobian, whether the Jacobian is nonsingular or singular. We develop such an approach and, based upon it, an efficient tensor method for solving large sparse systems of nonlinear equations. Test results indicate that this tensor method is significantly more efficient and robust than an efficient sparse Newton-based method, in terms of iterations, function evaluations, and execution time. © 1998 The Mathematical Programming Society, Inc. Published by Elsevier Science B.V.Work supported by the Mathematical, Information, and Computational Sciences Division subprogram of the Office of Computational and Technology Research, US Department of Energy, under Contract W-31-109-Eng-38, by the National Aerospace Agency under Purchase Order L25935D, and by the National Science Foundation, through the Center for Research on Parallel Computation, under Cooperative Agreement No. CCR-9120008.Research supported by AFOSR Grants No. AFOSR-90-0109 and F49620-94-1-0101, ARO Grants No. DAAL03-91-G-0151 and DAAH04-94-G-0228, and NSF Grant No. CCR-9101795.     

On combinatorial optimization problems on matroids with uncertain weights

In this paper the combinatorial optimization problem on weighted matroid is considered. It is assumed that the weights in the problem are ill-known and they are modeled as fuzzy intervals. The optimality of solutions and the optimality of elements are characterized. This characterization is performed in the setting of possibility theory. A method of choosing a solution under uncertainty is also proposed.     

Determination of eigenvalues of real symmetric para-p diagonal matrices

A method is presented for an accurate numerical determination of eigenvalues of real symmetric para-p diagonal matrices. The method takes advantage of the band structure to break up the matrix intop ×p blocks and performing algebraic operations including inversions on these blocks only, no matter what the size of the matrix is. The eigenvalues are determined independently one at a time. Thus any error in the determination of one eigenvalue does not affect the other eigenvalues. The method is ideally suited for the Schrödinger eigen alue problem of the anharmonic potentials, which is taken up in the following paper.     

一个t—范数下的不分明群问题

本文指出了不分明群中一个重要定理并不完全成立，从而将引出一系列理论问题。     

A general model for the release of active agents incorporated in swellable polymeric matrices

A theoretical model, which takes proper account of the simultaneous uptake of a liquid leachant by a polymer matrix and the consequent release of a bioactive or other solute incorporated therein, is presented and the kinetics of release predicted under various conditions is investigated. This model is shown to be both more rigorous and much more widely applicable than previous attempts to approach the above problem. Furthermore, it can be easily modified or extended, in accordance with the information available about any particular system. It is, therefore, expected that the model introduced here will prove useful as a basis for the design of monolithic controlled-release devices of this type or for the evaluation of the leachability of low-level and medium-level radioactive wastes “immobilized” in polymeric matrices.     

Hankel matrices and quadratic forms

A characterization of finite Hankei matrices is given and it is shown that such matrices arise naturally as matrix representations of scaled trace forms of field extensions and etale algebras. An algorithm is given for calculating the signature and the Hasse invariant of these scaled trace forms.     

The Application of Intelligent Technology on a Multi-Variable Dynamical System

Conventional control systems express control solutions by means of control expressions, usually mathematically based. To completely express the control solution, a vast amount of data is required. This is either difficult or virtually impossible to obtain. In contrast, knowledge-based solutions require far less plant data and mathematical expression. This reduces development time proportionately. A dynamical temperature control system is used in this project, within which a Fuzzy Logic Controller is used to maintain the system temperature. A Fuzzy Logic Control algorithm is implemented to test its performance in temperature control. An experimental analysis indicated that for approximately 50% decrease in flow rate, considering that the flow rate could not be increased to the desired value, owing to process problems, neither of the two temperature valves was subjected to oscillations. After approximately 30 s, the process temperature began to steady out. Mathematical reasoning would have increased the settling time tremendously.     

Generalized Lyapunov exponents in high-dimensional chaotic dynamics and products of large random matrices

We study the behavior of the generalized Lyapunov exponents for chaotic symplectic dynamical systems and products of random matrices in the limit of large dimensionsD. For products of random matrices without any particular structure the generalized Lyapunov exponents become equal in this limit and the value of one of the generalized Lyapunov exponents is obtained by simple arguments. On the contrary, for random symplectic matrices with peculiar structures and for chaotic symplectic maps the generalized Lyapunov exponents remains different forD , indicating that high dimensionality cannot always destroy intermittency.     

Renormalizing rectangles and other topics in random matrix theory

We consider random Hermitian matrices made of complex or realM×N rectangular blocks, where the blocks are drawn from various ensembles. These matrices haveN pairs of opposite real nonvanishing eigenvalues, as well asM–N zero eigenvalues (forM>N). These zero eigenvalues are kinematical in the sense that they are independent of randomness. We study the eigenvalue distribution of these matrices to leading order in the large-N, M limit in which the rectangularityr=M/N is held fixed. We apply a variety of methods in our study. We study Gaussian ensembles by a simple diagrammatic method, by the Dyson gas approach, and by a generalization of the Kazakov method. These methods make use of the invariance of such ensembles under the action of symmetry groups. The more complicated Wigner ensemble, which does not enjoy such symmetry properties, is studied by large-N renormalization techniques. In addition to the kinematical -function spike in the eigenvalue density which corresponds to zero eigenvalues, we find for both types of ensembles that if |r–1| is held fixed asN, theN nonzero eigenvalues give rise to two separated lobes that are located symmetrically with respect to the origin. This separation arises because the nonzero eigenvalues are repelled macroscopically from the origin. Finally, we study the oscillatory behavior of the eigenvalue distribution near the endpoints of the lobes, a behavior governed by Airy functions. Asr1 the lobes come closer, and the Airy oscillatory behavior near the endpoints that are close to zero breaks down. We interpret this breakdown as a signal thatr1 drives a crossover to the oscillation governed by Bessel functions near the origin for matrices made of square blocks.