求解一类约束优化信赖域方法的子问题

１引言设Ｈ为一给定的ｎ×ｎ对称矩阵,ｃＲ＂,本文考虑如｝的约束优化问题这里ａ＞０为给定的参数,Ｃ＝｛ｘＲｎｘ＜ａ是Ｒ”中的一个球体,Ｋ是一个简单凸闭集．当Ｋ＝Ｒｎ时,问题（Ｐ）便是无约束优化的信赖域子问题．当Ｋ＝｛ｘＲｎμ≤ｘ≤υ５,（μ１,μ２,…,μｎ）Ｔ,υ＝（υ１,υ２…,υｎ）Ｔ,且—∞＜μｉ＜υｉ＜ｖ＜＋∞,ｉ＝１,２,…,ｎ时,问题（Ｐ）便是用信赖域方法求解带上下界约束的优化问题时遇到的子问题．对于无约束信赖域方法的子问题已经有了比较成熟的算法［８,１２－１３,１５－１６］．Ｋ＝Ｒ…     

A characterization of almost simple K 3-groups

It is a well-known fact that characters of a finite group can give important information about the group's structure. Also it was proved by the third author of this article that a finite simple group can be uniquely determined by its character table. Here the authors attempt to investigate how to characterize a finite almost simple group by using less information of its character table, and successfully characterize the almost simple K₃-groups by their orders and at most three irreducible character degrees of their character tables.     

Bounded Tiles in ℚp are Compact Open Sets

Any bounded tile of the field Qp of p-adic numbers is a compact open set up to a set of zero Haar measure.In this note,we present two simple and direct proofs of this fact.     

Cauchy Matrices in the Observation of Diffusion Equations

Observability Gramians of diffusion equations have been recently connected to infinite Pick and Cauchy matrices. In fact, inverse or observability inequalities can be obtained after estimating the extreme eigenvalues of these structured matrices,with respect to the diffusion semi-group matrix. The purpose is hence to conduct a spectral study of a subclass of symmetric Cauchy matrices and present an algebraic way to show the desired observability results. We revisit observability inequalities for three different observation problems of the diffusion equation and show how they can be (re)stated through simple proofs.     

Why can we not make a perfect map?

The purpose of this note is to present a simple proof (without using the Gauss egregium theorem) of the following fact: To make a length preserving projection of the Earth is impossible.     

Zero-clusters of polynomials: Best approach in supercomputing era

Computing a zero-cluster of a polynomial sufficiently accurately within the available precision of computation has been an important issue from time immemorial. All the deterministic numerical methods so far known to us produce varying degree of errors. Often the errors are so dominant that the distinction between two zeros in the cluster becomes meaningfully difficult. Multiple zeros on the other hand can be more easily tackled and do not pose any serious computational problem. We discuss here the limits of both deterministic and randomized methods for zero-clusters and propose a simple exhaustive search algorithm that would obtain the zeros in a real/complex zero-cluster in a reasonable time. We present the computational error and computational/time complexity of this algorithm focusing on the fact that no measuring device can usually measure a quantity with an accuracy greater than 0.005%. We stress the fact that no other algorithm can perform better than the proposed algorithm in an ultra-high speed computing environment for most real-world problems.     

A class of Iteration functions for improving,simultaneously, approximations to the zeros of a polynomial

Approximations to all the zeros of a polynomial may be used to find a better approximation to one of the zeros. Using this fact, we present a simple approach to the derivation of a class of Iteration Functions for simultaneously improving approximations to zeros of a polynomial. Convergence properties are studied and computational results are included.     

On the calculation of the spectra of directionally perturbed operators

A class of operators with special spectral properties is defined. An operator in this class is fairly simple, acts in a separable Hilbert space, and can be perturbed so that an a priori given function from its domain is an eigenfunction of the perturbed operator. This fact is shown to be useful for constructing operators in mathematical physics. Specific examples are given.     

DISJOINT SUPERCYCLIC WEIGHTED PSEUDO-SHIFTS ON BANACH SEQUENCE SPACES

In this article, we present several equivalent conditions ensuring the disjoint supercyclicity of finite weighted pseudo-shifts acting on an arbitrary Banach sequence space.The disjoint supercyclic properties of weighted translations on locally compact discrete groups,the direct sums of finite classical weighted backward shifts, and the bilateral backward operator weighted shifts can be viewed as special cases of our main results. Furthermore, we exhibit an interesting fact that any finite bilateral weighted backward shifts on the space ?~2(Z) never satisfy the d-Supercyclicity Criterion by a simple proof.     

Trivializable rings of sets

In the present paper, we examine the ways in which rings of sets can be generated by as simple systems as possible. We investigate rings of sets which can be generated by totally ordered subsystems. All totally ordered systems which generate the same ring are, in essence, isomorphic. This fact leads to an order-theoretic classification of rings of sets. Every group-valued mapping defined on a totally ordered system of sets can be uniquely extended to an additive mapping on the ring of sets generated by the system. Two rings of sets are of the same order type if and only if the corresponding spaces of measures are isomorphic.     

Techniques for bounding the convergence rate of genetic algorithms

The main purpose of the present paper is the study of computational aspects, and primarily the convergence rate, of genetic algorithms (GAs). Despite the fact that such algorithms are widely used in practice, little is known so far about their theoretical properties, and in particular about their long‐term behavior. This situation is perhaps not too surprising, given the inherent hardness of analyzing nonlinear dynamical systems, and the complexity of the problems to which GAs are usually applied. In the present paper we concentrate on a number of very simple and natural systems of this sort, and show that at least for these systems the analysis can be properly carried out. Various properties and tight quantitative bounds on the long‐term behavior of such systems are established. It is our hope that the techniques developed for analyzing these simple systems prove to be applicable to a wider range of genetic algorithms, and contribute to the development of the mathematical foundations of this promising optimization method. ©1999 John Wiley & Sons, Inc. Random Struct. Alg., 14, 111–138, 1999     

Strong reality of finite simple groups

The classification of finite simple strongly real groups is complete. It is easy to see that strong reality for every nonabelian finite simple group is equivalent to the fact that each element can be written as a product of two involutions. We thus obtain a solution to Problem 14.82 of the Kourovka Notebook from the classification of finite simple strongly real groups.     

Optimal Feedback Control for Linear Systems with Input Delays Revisited

The design problem of optimal feedback control for linear systems with input delays is very important in many engineering applications. Usually, the linear systems with input delays are firstly converted into linear systems without delays, and then all the design procedures are based on the delay-free linear systems. In this way, the feedback controllers are not designed in terms of the original states. This paper presents some new closed-form formula in terms of the original states for the delayed optimal feedback control of linear systems with input delays. We firstly reveal the essential role of the input delay in the optimal control design of the linear system with a single input delay: the input delay postpones the action of the optimal control only. Based on this fact, we calculate the delayed optimal control and find that the optimal state can be represented by a simple closed-form formula, so that the delayed optimal feedback control can be obtained in a simple way. We show that the delayed feedback gain matrix can be “smaller” than that for the controlled system with zero input delay, which implies that the input delay can be considered as a positive factor. In addition, we give a general formula for the delayed optimal feedback control of time-variant linear systems with multiple input delays. To show the effectiveness and advantages of the main results, we present five illustrative examples with detailed numerical simulation and comparison.     

On the Variability of Demand for Families of Items

For the purposes of inventory control and production planning of multi-product lines, often consisting of hundreds or thousands of items differing only in minor ways from each other, it is often found that the variability of demand for individual items can be expressed as a simple function of the expected demand for each item. Usually this is not Poisson variability. In this paper a simple demand model is proposed which generates variability of the type often observed, and which in fact leads directly to the Quadratic Law of Burgin and Wild. It is noted how, in practice, it may be hard to distinguish between this law and the Power Law (Brown). The difference, however, is by no means unimportant, and this is discussed. The statistical problems involved in estimating either law from practical data are also shown to be non-trivial.     

Saddle-Point Optimality: A Look Beyond Convexity

The fact that two disjoint convex sets can be separated by a plane has a tremendous impact on optimization theory and its applications. We begin the paper by illustrating this fact in convex and partly convex programming. Then we look beyond convexity and study general nonlinear programs with twice continuously differentiable functions. Using a parametric extension of the Liu-Floudas transformation, we show that every such program can be identified as a relatively simple structurally stable convex model. This means that one can study general nonlinear programs with twice continuously differentiable functions using only linear programming, convex programming, and the inter-relationship between the two. In particular, it follows that globally optimal solutions of such general programs are the limit points of optimal solutions of convex programs.     

A new variational characterization of Jacobi fields along geodesies

A classical result of Riemannian geometry states that Jacobi fields along geodesics of a Riemannian manifold (Q, g) can be obtained as geodesies of the so-called «complete lift» of the metric g itself to the tangent bundle TQ. We show that this classical result is in fact a very simple consequence of a completely general theorem of Calculus of Variations.     

Simplicity of the branching of representations of the groups GL(<Emphasis Type="Italic">n,q</Emphasis>) under the parabolic restriction: An elementary proof

It is well known that the branching of representations of the groups GL(n, q) under the parabolic restriction is simple. Apparently, an elementary proof of this important fact has not been found so far. We present such a proof, which uses a method not standard for the representation theory of finite groups.     

Modelling nonlinear sequence generators in terms of linear cellular automata

In this work, a wide family of LFSR-based sequence generators, the so-called clock-controlled shrinking generators (CCSGs), has been analyzed and identified with a subset of linear cellular automata (CA). In fact, a pair of linear models describing the behavior of the CCSGs can be derived. The algorithm that converts a given CCSG into a CA-based linear model is very simple and can be applied to CCSGs in a range of practical interest. The linearity of these cellular models can be advantageously used in two different ways: (a) for the analysis and/or cryptanalysis of the CCSGs and (b) for the reconstruction of the output sequence obtained from this kind of generators.     

Approximation of centroid end-points and switch points for replacing type reduction algorithms

Despite several years of research, type reduction (TR) operation in interval type-2 fuzzy logic system (IT2FLS) cannot perform as fast as a type-1 defuzzifier. In particular, widely used Karnik–Mendel (KM) TR algorithm is computationally much more demanding than alternative TR approaches. In this work, a data driven framework is proposed to quickly, yet accurately, estimate the output of the KM TR algorithm using simple regression models. Comprehensive simulation performed in this study shows that the centroid end-points of KM algorithm can be approximated with a mean absolute percentage error as low as 0.4%. Also, switch point prediction accuracy can be as high as 100%. In conjunction with the fact that simple regression model can be trained with data generated using exhaustive defuzzification method, this work shows the potential of proposed method to provide highly accurate, yet extremely fast, TR approximation method. Speed of the proposed method should theoretically outperform all available TR methods while keeping the uncertainty information intact in the process.     

Shallow, deep and very deep cuts in the analytic center cutting plane method

In this paper we show that the cut does not need to go through the query point: it can be deep or shallow. The primal framework leads to a simple analysis of the potential variation, which shows that the inequality needed for convergence of the algorithm is in fact attained at the first iterate of the feasibility step. Received July 3, 1996 / Revised version received July 11, 1997 Published online August 18, 1998