Stability of solutions for a class of nonlinear cone constrained optimization problems,part 2: Application to parameter estimation

In this paper we study a problem of parameter estimation in two point boundary value problems. Using a stability theorem for nonlinear cone constrained optimization problems derived in Part 1 of this paper we investigate stability properties of the solutions of the parameter estimation problem in the output-least-squares formulation.     

Regularization of nonlinear illposed problems with closed operators

In this paper Tikhonov regularization for nonlinear illposed problems is investigated. The regularization term is characterized by a closed linear operator, permitting seminorm regularization in applications. Results for existence, stability, convergence and con- vergence rates of the solution of the regularized problem in terms of the noise level are given. An illustrating example involving parameter estimation for a one dimensional stationary heat equation is given.     

On box schemes for elartliptic variational inequalities

General finite volume approximations in two and three space dimensions are studied for the discretization of interior and boundary obstacle problems with mixed boundary conditions. First order convergence of the finite volume (box) solution is shown in the energy norm. Based on a discrete maximum principle there are proposed two penalization techniques for the solution of the finite volume inequalities. By coupling of discretization and penalty parameters the overall error is analyzed. The iterative solution of the penalty problems is discussed. Finally, results of numerical experiments are presented to illustrate the convergence behaviour between the exact, the box and the penalty solutions.     

The lagrange-newton method for infinite-dimensional optimization problems

This paper investigates local convergence properties of the Lagrange-Newton method for optimization problems in reflexive Banach spaces. Sufficient conditions for quadratic convergence of optimal solutions and Lagrange multipliers are given. The results are applied to optimal control problems.     

Enlargement procedure for resolution of singularities at simple singular solutions of nonlinear equations

Summary A solution of a nonlinear equation in Hilbert spaces is said to be a simple singular solution if the Fréchet derivative at the solution has one-dimensional kernel and cokernel. In this paper we present the enlargement procedure for resolution of singularities at simple singular solutions of nonlinear equations. Once singularities are resolved, we can compute accurately the singular solution by Newton's method. Conditions for which the procedure terminates in finite steps are given. In particular, if the equation defined in ⁿ is analytic and the simple singular solution is geometrically isolated, the procedure stops in finite steps, and we obtain the enlarged problem with an isolated solution. Numerical examples are given.This research is partially supported by Grant-in-Aid for Encouragment of Young Scientist No. 60740119, the Ministry of EducationDedicated to Professor Seiiti Huzino on his 60th birthday     

Nonlinear hybrid procedures and fixed point iterations

Let (x_n) and (?_n) be two vector sequences converging to a common limit. First, we shall define nonlinear hybrid procedures which consist of constructing a new vector sequence (y_n) with better convergence properties than (x_n) and (?_n). Then, this procedure is used for accelerating the convergence of a given sequence and applied to the construction of fixed point methods. New methods are derived. Finally, the connection between fixed point iterations and methods for the numerical integration of differential equations is also exploited. Numerical results are given.     

Ein selektives Projektions-Iterationsverfahren und Anwendungen auf Verzweigungsprobleme

Summary In the present paper a projection-iteration procedure for solving nonlinear operator-equations is described in abstract spaces. The method possesses a selection property by yielding only those solutions which have certain stability properties. Parameter-dependent equations and bifurcation problems are considered as well. It is shown that in the case of bifurcation near a simple eigenvalue, the method converges locally always to the stable solution except for initial values on a proper submanifold. Numerical results for a nonlinear boundary-value problem are given illustrating this selection character.

Diese Arbeit wurde von der Deutschen Forschungsgemeinschaft unter Ki 131/2 gefördert     

Hilbert's theorem 90 in matrix rings †

The fact that every unitary matrix can be expressed in the form A^-1A^? (observed by C. R. Johnson) is interpreted as a "Hilbert Theorem 90" in rings with an involution. Several observations connected with this are included.     

Asymptotic behaviour of the tandem queueing system with identical service times at both queues

Consider a tandem queue consisting of two single-server queues in series, with a Poisson arrival process at the first queue and arbitrarily distributed service times, which for any customer are identical in both queues. For this tandem queue, we relate the tail behaviour of the sojourn time distribution and the workload distribution at the second queue to that of the (residual) service time distribution. As a by-result, we prove that both the sojourn time distribution and the workload distribution at the second queue are regularly varying at infinity of index 1−ν, if the service time distribution is regularly varying at infinity of index −ν (ν>1). Furthermore, in the latter case we derive a heavy-traffic limit theorem for the sojourn time S ⁽²⁾ at the second queue when the traffic load ρ↑ 1. It states that, for a particular contraction factor Δ (ρ), the contracted sojourn time Δ (ρ) S ⁽²⁾ converges in distribution to the limit distribution H(·) as ρ↑ 1 where .     

Hilbert's theorem 90 in matrix rings

The fact that every unitary matrix can be expressed in the form A^-1A^* (observed by C. R. Johnson) is interpreted as a "Hilbert Theorem 90" in rings with an involution. Several observations connected with this are included.     

Some supplementary results on the 1+\sqrt 2 order method for the solution of nonlinear equations

Summary Recently an iterative method for the solution of systems of nonlinear equations having at leastR-order 1+ for simple roots has been investigated by the author [7]; this method uses as many function evaluations per step as the classical Newton method. In the present note we deal with several properties of the method such as monotone convergence, asymptotic inclusion of the solution and convergence in the case of multiple roots.     

On 2‐Colorings of Hypergraphs

In this article, we continue the study of 2‐colorings in hypergraphs. A hypergraph is 2‐colorable if there is a 2‐coloring of the vertices with no monochromatic hyperedge. Let denote the class of all k‐uniform k‐regular hypergraphs. It is known (see Alon and Bregman [Graphs Combin. 4 (1988) 303–306] and Thomassen [J. Amer. Math. Soc. 5 (1992), 217–229] that every hypergraph is 2‐colorable, provided . As remarked by Alon and Bregman the result is not true when , as may be seen by considering the Fano plane. Indeed there are several constructions for building infinite families of hypergraphs in that are not 2‐colorable. Our main result in this paper is a strengthening of the above results. For this purpose, we define a set X of vertices in a hypergraph H to be a free set in H if we can 2‐color such that every edge in H receives at least one vertex of each color. We conjecture that for , every hypergraph has a free set of size in H. We show that the bound cannot be improved for any and we prove our conjecture when . Our proofs use results from areas such as transversal in hypergraphs, cycles in digraphs, and probabilistic arguments.     

On smoothness and invariance properties of the gauss-newton method

We consider systems of m nonlinear equations in m + p unknowns which have p-dimensional solution manifolds. It is well-known that the Gauss-Newton method converges locally and quadratically to regular points on this manifold. We investigate in detail the mapping which transfers the starting point to its limit on the manifold. This mapping is shown to be smooth of one order less than the given system. Moreover, we find that the Gauss-Newton method induces a foliation of the neighborhood of the manifold into smooth submanifolds. These submanifolds are of dimension m, they are invariant under the Gauss-Newton iteration, and they have orthogonal intersections with the solution manifold.     

Variable selection in model-based discriminant analysis

A general methodology for selecting predictors for Gaussian generative classification models is presented. The problem is regarded as a model selection problem. Three different roles for each possible predictor are considered: a variable can be a relevant classification predictor or not, and the irrelevant classification variables can be linearly dependent on a part of the relevant predictors or independent variables. This variable selection model was inspired by a previous work on variable selection in model-based clustering. A BIC-like model selection criterion is proposed. It is optimized through two embedded forward stepwise variable selection algorithms for classification and linear regression. The model identifiability and the consistency of the variable selection criterion are proved. Numerical experiments on simulated and real data sets illustrate the interest of this variable selection methodology. In particular, it is shown that this well ground variable selection model can be of great interest to improve the classification performance of the quadratic discriminant analysis in a high dimension context.     

Nonlinear stability and convergence of finite-difference methods for the “good” Boussinesq equation

Summary The good Boussinesq equationu _tt =–u _xxxx +u _xx +(u ²) _xx has recently been found to possess an interesting soliton-interaction mechanism. In this paper we study the nonlinear stability and the convergence of some simple finite-difference schemes for the numerical solution of problems involving the good Boussinesq equation. Numerical experimentas are also reported.     

Existence of Five MOLS of Orders 18 and 60

In this article, we provide direct constructions for five mutually orthogonal Latin squares (MOLS) of orders and 60. For , these come from a new (60, 6, 1) difference matrix. For , the required construction is obtained by combining two different methods that were used in the constructions of four MOLS(14) and eight MOLS(36).     

Numerical computation of nonsimple turning points and cusps

Summary A method is developed for the numerical computation of a double turning point corresponding to a cusp catastrophe of a nonlinear operator equation depending on two parameters. An augmented system containing the original equation is introduced, for which the cusp point is an isolated solution. An efficient implementation of Newton's method in the finite-dimensional case is presented. Results are given for some chemical engineering problems and this direct method is compared with some other techniques to locate cusp points.     

Parallel nonlinear multisplitting methods

Summary Linear multisplitting methods are known as parallel iterative methods for solving a linear systemAx=b. We extend the idea of multisplittings to the problem of solving a nonlinear system of equationsF(x)=0. Our nonlinear multisplittings are based on several nonlinear splittings of the functionF. In a parallel computing environment, each processor would have to calculate the exact solution of an individual nonlinear system belonging to his nonlinear multisplitting and these solutions are combined to yield the next iterate. Although the individual systems are usually much less involved than the original system, the exact solutions will in general not be available. Therefore, we consider important variants where the exact solutions of the individual systems are approximated by some standard method such as Newton's method. Several methods proposed in literature may be regarded as special nonlinear multisplitting methods. As an application of our systematic approach we present a local convergence analysis of the nonlinear multisplitting methods and their variants. One result is that the local convergence of these methods is determined by an induced linear multisplitting of the Jacobian ofF.Dedicated to the memory of Peter Henrici