Reduced-order filtering with energy-to-peak performance for discrete-time Markovian jumping systems

In this paper, the l₂–l (energy-to-peak) performanceof the discrete-time Markovian jump linear system is investigated.The jump parameters are modelled by a discrete-time Markov process.Furthermore, we study the l₂–l reduced-order filteringproblem for the Markovian jump linear system. A reduced-orderfilter with the same randomly jumping parameters is proposedwhich can make the error systems with Markovian jump parametersstochastically stable with a prescribed l₂–lperformance.Sufficient conditions in terms of linear matrix inequalities(LMIs) and a coupling non-convex rank constraint are derivedfor the existence of a solution to the reduced-order filteringproblems. A numerical example is given to illustrate the designprocedures.     

Error Bounds for the Truncation of Infinite Linear Differential Systems

An infinite system of linear differential equations of the form = Ax + f, x(0) = y is considered, wherex, y and f are infinite column vectors in E, and A is a constantinfinite matrix defining a bounded operator on E, where E isl₁, (c₀), or l. Explicit error bounds are obtained for the approximationof the solution of the infinite system by the solutions of thefinite truncated systems.     

The Minimax Solution of Linear Equations Subject to Linear Constraints

This paper presents an efficient active set method for computingthe l_m solution of linear equations subject to linear constraints.The search direction of the proposed method is obtained viathe solution of a linear least-squares problem. It is shownthat the steepest-descent direction can be obtained by solvingthe same least-squares problem with simple bounds on the variables.This direction is used to prevent cycling at degenerate deadpoints. Numerical experiments illustrate the feasibility ofthe new method.     

Conditions for Superlinear Convergence in l1 and l{infty} Solutions of Overdetermined Non-linear Equations

Let n be the number of unknowns in an overdetermined systemof non-linear equations. If fewer than n equations are satisfiedin an l₁ solution or if fewer than (n+1) maximum residuals occurin an lsolution, then iterative methods of calculation convergesuperlinearly only if some second derivative information isused. This paper establishes some conditions on second derivativeestimates that are necessary and sufficient for superlinearconvergence.     

The Convergence of a Levenberg–Marquardt Method for Nonlinear Inequalities

In this paper, we consider the least l ₂-norm solution for a possibly inconsistent system of nonlinear inequalities. The objective function of the problem is only first-order continuously differentiable. By introducing a new smoothing function, the problem is approximated by a family of parameterized optimization problems with twice continuously differentiable objective functions. Then a Levenberg–Marquardt algorithm is proposed to solve the parameterized smooth optimization problems. It is proved that the algorithm either terminates finitely at a solution of the original inequality problem or generates an infinite sequence. In the latter case, the infinite sequence converges to a least l ₂-norm solution of the inequality problem. The local quadratic convergence of the algorithm was produced under some conditions.     

An Augmented Galerkin Method for First Kind Fredholm Equations

We describe an augmented Galerkin technique for the numericalsolution of first kind Fredholm equations, which is simple touse and which has the considerable advantage of providing acheaply computed numerical criterion for the existence of asolution of the equations under study. The method has guaranteedstability, and leads to a standard linear programming problem(when posed in the l₁ or l norms). It is much faster than themethod recommended in a recent review by Lewis (1975); numericalcomparisons indicate that it achieves comparable accuracy. Theexistence criterion also appears effective in practice.     

Mahler Measure of Alexander Polynomials

Let l be an oriented link of d components in a homology 3-sphere.For any nonnegative integer q, let l(q) be the link of d–1components obtained from l by performing 1/q surgery on itsdth component l_d. The Mahler measure of the multivariable Alexanderpolynomial _l(q) converges to the Mahler measure of _l as q goesto infinity, provided that l_d has nonzero linking number withsome other component. If l_d has zero linking number with eachof the other components, then the Mahler measure of _l(q) hasa well defined but different limiting behavior. Examples aregiven of links l such that the Mahler measure of _l is small.Possible connections with hyperbolic volume are discussed.     

The global linear and local quadratic convergence of a non-interior continuation algorithm for the LCP

^** Email: zhenghaihuang{at}yahoo.com.cn; huangzhenghai{at}hotmail.com In this paper, we propose a non-interior continuation algorithmfor solving the P₀-matrix linear complementarity problem (LCP),which is conceptually simpler than most existing non-interiorcontinuation algorithms in the sense that the proposed algorithmonly needs to solve at most one linear system of equations ateach iteration. We show that the proposed algorithm is globallyconvergent under a common assumption. In particular, we showthat the proposed algorithm is globally linearly and locallyquadratically convergent under some assumptions which are weakerthan those required in many existing non-interior continuationalgorithms. It should be pointed out that the assumptions usedin our analysis of both global linear and local quadratic convergencedo not imply the uniqueness of the solution to the LCP concerned.To the best of our knowledge, such a convergence result hasnot been reported in the literature.     

A Numerical Method for the Solution of the Double Eigenvalue Problem

A generalization of the Rayleigh quotient iterative method,called the Minimum Residual Quotient Iteration (MRQI), is derivedfor the numerical solution of the 2-parameter eigenvalue problem;i.e. to find scalars µ and a corresponding vector x satisfyingthe following equations, Ax = B₁x + µB₂x, ||x|| = 1, f(x) = 0, where A and B are nxn real matrices, ||.|| denotes the l₂ normand f is a real functional. The method is applied to doubleeigenvalue problems for ordinary differential equations andcomputational results are presented.     

Uniform Stability for Time-Varying Infinite-Dimensional Discrete Linear Systems

Stability for time-varying discrete linear systems in a Banachspace is investigated. On the one hand is established a fairlycomplete collection of necessary and sufficient conditions foruniform asymptotic equistability for input-free systems. Thisincludes uniform and strong power equistability, and uniformand strong l_p-equistability, among other technical conditionswhich also play an essential role in stability theory. On theother hand, it is shown that uniform asymptotic equistabilityfor input-free systems is equivalent to each of the followingconcepts of uniform stability for forced systems: l_p-input l_p-state,e_o-input e_o-state, bounded-input bounded-state, l_p-input bounded-state(with p>1), e_o-input bounded-state, and convergent-inputbounded-state; these are also equivalent to their nonuniformcounterparts. For time-varying convergent systems, the aboveis also equivalent to convergent-input convergent-state stability.The proofs presented here are all ‘lementary’ inthe sense that they are based essentially only on the Banach–Steinhaustheorem.     

Algebraic decay and variable speeds in wavefront solutions of a scalar reaction-diffusion equation

Wavefront solutions of scalar reaction-diffusion equations havebeen intensively studied for many years. There are two basiccases, typified by quadratic and cubic kinetics. An intermediatecase is considered in this paper, namely, u_l = u_xx + u²(1 –u). It is shown that there is a unique travelling-wave solution,with a speed 1/2, for which the decay to zero ahead of the waveis exponential with x. Moreover, numerical evidence is presentedwhich suggests that initial conditions with such exponentialdecay evolve to this travelling-wave solution, independentlyof the half-life of the initial decay. It is then shown thatfor all speeds greater than 1/2 there is also a travelling-wavesolution, but that these faster waves decay to zero algebraically,in proportion to 1/x. The numerical evidence suggests that initialconditions with such a decay rate evolve to one of these fasterwaves; the particular speed depends in a simple way on the detailsof the initial condition. Finally, initial conditions decayingalgebraically for a power law other than 1/x are considered.It is shown numerically that such initial conditions evolveeither to an algebraically decaying travelling wave, or in somecases to a wavefront whose shape and speed vary as a functionof time. This variation is monotonic and can be quite pronounced,and the speed is a function of u as well as of time. Using asimple linearization argument, an approximate formula is derivedfor the wave speed which compares extremely well with the numericalresults. Finally, the extension of the results to the more generalcase of u_l = u_xx + u^m(1 – u), with m > 1, is discussed.     

On a Question of Haskell P. Rosenthal Concerning a Characterization of c0 and lp

The following property of a normalized basis in a Banach spaceis considered: any normalized block sequence of the basis hasa subsequence equivalent to the basis. Under uniformity or othernatural assumptions, a basis with this property is equivalentto the unit vector basis of c₀ or l_p. An analogous problem concerningspreading models is also addressed. 2000 Mathematics SubjectClassification 46B20 (primary), 46B15 (secondary).     

On the Poles of Igusa's Local Zeta Function for Algebraic Sets

Let K be a p-adic field, let Z (s, f), sC, with Re(s) > 0,be the Igusa local zeta function associated to f(x) = (f₁(x),..., f_l(x)) [K (x₁, ..., x_n)]^l, and let be a Schwartz–Bruhatfunction. The aim of this paper is to describe explicitly thepoles of the meromorphic continuation of Z (s, f). Using resolutionof singularities it is possible to express Z (s, f) as a finitesum of p-adic monomial integrals. These monomial integrals arecomputed explicitly by using techniques of toroidal geometry.In this way, an explicit list of the candidates for poles ofZ (s, f) is obtained. 2000 Mathematics Subject Classification11S40, 14M25, 11D79.     

Bounds on Norms of Compound Matrices and on Products of Eigenvalues

An upper bound on operator norms of compound matrices is presented,and special cases that involve the l₁, l₂ and l norms are investigated.The results are then used to obtain bounds on products of thelargest or smallest eigenvalues of a matrix. 1991 MathematicsSubject Classification 15A15, 15A18, 15A42.     

Discrete l1 Approximation by Rational Functions

The problem of finding a best approximation by a rational functionto discrete data, using the l₁ norm, is considered. An algorithmis developed which is frequently convergent in a finite numberof steps, and failing this usually has a second-order convergencerate. Details are given of the application of the algorithmto a number of rational approximation problems.     

(k, l)-Algebraic Stability of Gauss Methods

It is well known that the s-stage Gauss Runge-Kutta methodsof order 2s are algebraically stable, or equivalently (1, 0)-algebraicallystable. In this paper, we show that there exists some l_s >0 such that the Gauss methods are (k, l) algebraically stablefor l [0, l_s) with k(l)=e^2l+O(l^p+1, where p=2s if s=1 or s=2,and p=2 if s>3.     

Generalisation of the Bogomolov-Miyaoka-Yau Inequality to Singular Surfaces

The paper considers pairs (X, B) where X is a normal projectivesurface over C, and B is a Q-divisor whose coefficients are1 or 1–1/m for some natural number m. A log canonicalsingularity on such a pair is a quotient by a finite or infinitegroup, so if (X, B) has log canonical singularities, the orbifoldEuler number e_orb(X, B) can be defined. The main result is aBogomolov-Miyaoka-Yau-type inequality which implies that if(X, B) has log canonical singularities and (X, K_X + B) 0 then(K_X+B)² 3e_orb(X, B). The actual inequality proved is somewhatstronger and it also implies all the previously published versionsof the Bogomolov-Miyaoka-Yau inequality. The proof involvesthe Log Minimal Model Program, Q-sheaves when K_X+B is nef, anda study of the changes in the two sides of the inequality undera contraction. The paper also contains a further generalisationwhere the coefficients of B can be arbitrary rational numbersin [0, 1], a different condition is imposed on the singularitiesand K_X+B is required to be nef. Some applications of the inequalitiesare also given, for example, estimating the number of singularitiesor certain kinds of configurations of curves on surfaces. 1991Mathematics Subject Classification: 14J17, 14J60, 14C17.     

Semimartingales and geometric inequalities on manifolds

The purpose of this paper is twofold. First, we use resultson Jacobi fields to study the stochastic differential equations(SDEs) for exp_Xt( exp_Xt^-1(Y_t)) with specially constructed coupledsemimartingales X and Y on a complete, simply connected Riemannianmanifold M with constant sectional curvature. Secondly, we applythese SDEs to obtain an analogue for M of a result of Borellconcerning an inequality relating the solutions of the parabolicequation / t = 1/2 – h, with Dirichlet boundary condition,on three convex sets in Euclidean space. From the latter, therefollows an inequality involving the first eigenvalues of theLaplacian on those convex sets with the Dirichlet boundary condition,analogous to an inequality in Euclidean space which is equivalentto the Brunn–Minkowski inequality of these eigenvaluesobtained by Brascamp and Lieb.     

Algebras of Symmetric Holomorphic Functions on lp

The authors study the algebra of uniformly continuous holomorphicsymmetric functions on the ball of l_p, investigating in particularthe spectrum of such algebras. To do so, they examine the algebraof symmetric polynomials on l_p-spaces, as well as finitely generatedsymmetric algebras of holomorphic functions. Such symmetricpolynomials determine the points in l_p up to a permutation.     

One Cubic Diophantine Inequality

Suppose that G(x) is a form, or homogeneous polynomial, of odddegree d in s variables, with real coefficients. Schmidt [15]has shown that there exists a positive integer s₀(d), whichdepends only on the degree d, so that if s s₀(d), then thereis an x Z^s\{0} satisfying the inequality |G(x)|<1. (1) In other words, if there are enough variables, in terms of thedegree only, then there is a nontrivial solution to (1). Lets₀(d) be the minimum integer with the above property. In thecourse of proving this important result, Schmidt did not explicitlygive upper bounds for s₀(d). His methods do indicate how todo so, although not very efficiently. However, in fact muchearlier, Pitman [13] provided explicit bounds in the case whenG is a cubic. We consider a general cubic form F(x) with realcoefficients, in s variables, and look at the inequality |F(x)|<1. (2) Specifically, Pitman showed that if s(1314)²⁵⁶–1, (3) then inequality (2) is non-trivially soluble in integers. Wepresent the following improvement of this bound.