Optimal Finite Characterization of Linear Problems with Inexact Data

Abstract. For many linear problems, in order to check whether a certain property is true for all matrices A from an interval matrix A, it is sufficient to check this property for finitely many “vertex” matrices A ∈ A. J. Rohn has discovered that we do not need to use all 2ⁿ² vertex matrices, it is sufficient to only check these properties for 2²ⁿ⁻¹ ≪ 2ⁿ² vertex matrices of a special type A_yz. In this paper, we show that a further reduction is impossible: without checking all 2²ⁿ⁻¹ matrices A_yz, we cannot guarantee that the desired property holds for all A ϵ A. Thus, these special vertex matrices provide an optimal finite characterization of linear problems with inexact data.     

Existence of triple positive solutions of two-point right focal boundary value problems on time scales

We consider the following boundary value problem, (−1)ⁿ⁻¹y^Δn(t)=(−1)^p+1F(t,y(σⁿ⁻¹(t))),t[a,b]∩T, y^Δn(a)=0,0≤i≤p−1, y^Δn(σ(b))=0,p≤i≤n−1,where n ≥ 2, 1 ≤ p ≤ n - 1 is fixed and T is a time scale. By applying fixed-point theorems for operators on a cone, existence criteria are developed for triple positive solutions of the boundary value problem. We also include examples to illustrate the usefulness of the results obtained.     

An efficient architecture for designing reverse converters based on a general three-moduli set

In this paper, a high-speed, low-cost and efficient design of reverse converter for the general three-moduli set {2^α, 2^β − 1, 2^β + 1} where α < β is presented. The simple proposed architecture consists of a carry save adder (CSA) and a modulo adder. As a result it can be efficiently implemented in VLSI circuits. The values of α and β are set in order to provide the desired dynamic range and also to obtain a balanced moduli set. Based on the above, two new moduli sets {2^n+k, 2²ⁿ − 1, 2²ⁿ + 1} and {2²ⁿ⁻¹, 2²ⁿ⁺¹ − 1, 2²ⁿ⁺¹ + 1}, which are the special cases of the moduli set {2^α, 2^β − 1, 2^β + 1} are proposed. The reverse converters for these new moduli sets are derived from the proposed general architecture with better performance compared to the other reverse converters for moduli sets with similar dynamic range.     

Fast Parallel Algorithms for Solving Triangular Systems of Linear Equations on the Hypercube

This paper presents efficient hypercube algorithms for solving triangular systems of linear equations by using various matrix partitioning and mapping schemes. Recently, several parallel algorithms have been developed for this problem. In these algorithms, the triangular solver is treated as the second stage of Gauss elimination. Thus, the triangular matrix is distributed by columns (or rows) in a wrap fashion since it is likely that the matrix is distributed this way after an LU decomposition has been done on the matrix. However, the efficiency of the algorithms is low. Our motivation is to develop various data partitioning and mapping schemes for hypercube algorithms by treating the triangular solver as an independent problem. Theoretically, the computation time of our best algorithm is ((12p + 1)n² + 36p³ − 28p²)/(24p²), and an upper bound on the communication time is 2αp log p (log n − log p) + 2α(log n − log p − 1) log p + (cn/p − 2c)(2 log p − 1) + log p(cn − c − α), where α is the (communication startup time)/(one entry scanning time), c is a constant, n is the order of the triangular system and p is the number of nodes in the hypercube. Experimental results show that the algorithm is efficient. The efficiency of the algorithm is 0.945 when p = 2, n = 513, and 0.93 when p = 8, n = 1025.     

Existence of positive solutions for th-order singular superlinear boundary value problems

This paper investigates the existence of positive solutions for 2nth-order (n>1) singular superlinear boundary value problems. A necessary and sufficient condition for the existence of C²ⁿ⁻²[0,1] as well as C²ⁿ⁻¹[0,1] positive solutions is given by constructing a special cone and with the e-Norm.     

Empirical Bayes estimation of the scale parameter in a Pareto distribution

Let f(xθ) = αθ^αx^−(α+1)I(x>θ) be the pdf of a Pareto distribution with known shape parameter α>0, and unknown scale parameter θ. Let {(X_i, θ_i)} be a sequence of independent random pairs, where X_i's are independent with pdf f(xα_i), and θ_i are iid according to an unknown distribution G in a class of distributions whose supports are included in an interval (0, m), where m is a positive finite number. Under some assumption on the class and squared error loss, at (n + 1)th stage we construct a sequence of empirical Bayes estimators of θ_n+1 based on the past n independent observations X₁,…, X_n and the present observation X_n+1. This empirical Bayes estimator is shown to be asymptotically optimal with rate of convergence O(n^−1/2). It is also exhibited that this convergence rate cannot be improved beyond n^−1/2 for the priors in class .     

Embeddings of Star Graphs into Optical Meshes without Bends

We show an embedding of the star graph into a rectangular optical multichannel mesh ofddimensions such that the embedding has no bends; that is, neighbors in the star graph always differ in exactly one coordinate in the mesh, to facilitate one-hop optical communication. To embed ann-star, the mesh can have any number of dimensionsdbetween 1 andn− 1. The embedding has load 1 and an expansion of at mostn^{d − 1}/d!. The size of the mesh will be at most We optimize the size of the host mesh using clique-partitioning to produce embeddings with expansions as low as unity. In two dimensions, for evenn, the mesh will be no larger thann×n(n− 2)!, and have an expansion of no more than 1 1/(n− 1). Further, we show how we can use a contraction method to efficiently embed the star graph into an optical mesh with near-unity aspect ratios. Contraction on a two-dimensional embedding will yield a mesh of size no larger thann×nfor evennwith a load of (n− 2)!.     

Riccati equations and normalized coprime factorizations for strongly stabilizable infinite-dimensional systems

The first part of the paper concerns the existence of strongly stabilizing solutions to the standard algebraic Riccati equation for a class of infinite-dimensional systems of the form Σ(A,B,S^−1/2B^*,D), where A is dissipative and all the other operators are bounded. These systems are not exponentially stabilizable and so the standard theory is not applicable. The second part uses the Riccati equation results to give formulas for normalized coprime factorizations over H_∞ for positive real transfer functions of the form D+S^−1/2B^*(author−A)⁻¹,B.     

The number of limit cycles for a family of polynomial systems

In this paper, the number of limit cycles in a family of polynomial systems was studied by the bifurcation methods. With the help of a computer algebra system (e.g., Maple 7.0), we obtain that the least upper bound for the number of limit cycles appearing in a global bifurcation of systems (2.1) and (2.2) is 5n + 5 + (1 − (−1)ⁿ)/2 for c ≠ 0 and n for c ≡ 0.     

Characterizations of distributions of exponential or geometric type by the integrated lack of memory property and record values

Let be a sequence of i.i.d. random variables, the sequence of its upper record values (i.e. L(0) = 1, L(n) = inf{jX_j>X_L(n−1} for n≥1). Without any assumptions to the support of P^X₁ the equidistribution of X₁ and a record increment X_L(n − X_L(n−1), n ≥ 1 yields X₁ to be either exponentially or geometrically distributed according to whether the additive subgroup generated by the support of P^X₁ is dense or a lattice in . The integrated lack of memory property can easily be reduced to the above problem for the case n = 1. Similarly the independence of X_L(n−1) and X_L(n) − X_L(n−1) for some n>1 characterizes X₁ to have e exponential or a geometric tail provided that the support of P^X₁ is bounded to the left and its right extremity no atom. Hence, if also its left extremity is no atom the independence of X_L(n−1) and X_L(n−1) − X_L(n−1) characterizes X₁ to be exponentially distributed.     

On Cryptographic Propagation Criteria for Boolean Functions

We determine the functions on GF(2)ⁿ which satisfy the propagation criterion of degree n−2, PC(n−2). We study subsequently the propagation criterion of degree ℓ and order k and its extended version EPC. We determine those Boolean functions on GF(2)ⁿ which satisfy PC(ℓ) of order kn−ℓ−2. We show that none of them satisfies EPC(ℓ) of the same order. We finally give a general construction of nonquadratic functions satisfying EPC(ℓ) of order k. This construction uses the existence of nonlinear, systematic codes with good minimum distances and dual distances (e.g., Kerdock codes and Preparata codes).     

A further result on the oscillation of delay difference equations

In this paper, we are concerned with the delay difference equations of the form

(*)

y_n+1 − y_n + p_ny_n−k = 0, N = 0, 1, 2, …,

(*)where p_n ≥ 0 and k is a positive integer. We prove by using a new technique that

guarantees that all solutions of equation (*) oscillate, which improves many previous well-known results. In particular, our theorems also fit the case where Σⁿ⁻¹_i=n−kp_i ≤ k^k+1/(k + 1)^k+1. In addition, we present a nonoscillation sufficient condition for equation (*).     

The query complexity of order-finding

We consider the problem where π is an unknown permutation on {0,1,…,2ⁿ−1}, y₀{0,1,…,2ⁿ−1}, and the goal is to determine the minimum r>0 such that π^r(y₀)=1. Information about π is available only via queries that yield π^x(y) from any x{0,1,…,2^m−1} and y{0,1,…,2ⁿ−1} (where m is polynomial in n). The main resource under consideration is the number of these queries. We show that the number of queries necessary to solve the problem in the classical probabilistic bounded-error model is exponential in n. This contrasts sharply with the quantum bounded-error model, where a constant number of queries suffices.     

On the Construction of Classes of Suffix Trees for Square Matrices: Algorithms and Applications

We provide a uniform framework for the study of index data structures for a two-dimensional matrixTEXT[1:n, 1:n] whose entries are drawn from an ordered alphabetΣ. An index forTEXTcan be informally seen as the two-dimensional analog of the suffix tree for a string. It allows on-line searches and statistics to be performed onTEXTby representing compactly theΘ(n³) square submatrices ofTEXTin optimalO(n²) space. We identify 4ⁿ⁻¹families of indices forTEXT, each containing ∏ⁿ_i=1 (2i−1)! isomorphic data structures. We also develop techniques leading to a single algorithm that efficiently builds any index in any family inO(n² log n) time andO(n²) space. Such an algorithm improves in various respects the algorithms for the construction of the PAT tree and the Lsuffix tree. The framework and the algorithm easily generalize tod>2 dimensions. Moreover, as part of our algorithm, we provide new algorithmic tools that yield a space-efficient implementation of the “naming scheme” of R. Karpet al.(in“Proceedings, Fourth Symposium on Theory of Computing,” pp. 125–136) for strings and matrices.     

A generalization of concavity for finite differences

The concept of concavity is generalized to discrete functions, u, satisfying the n^th-order difference inequality, (−1)^n−kΔⁿu(m) ≥ 0, M = 0, 1,..., N and the homogeneous boundary conditions, u(0) = … = u(k−1) = 0, u(N + k + 1) = … = u(N + n) = 0 for some k “1, …, n − 1”. A piecewise polynomial is constructed which bounds u below. The piecewise polynomial is employed to obtain a positive lower bound on u(m) for m = k, …, N + k, where the lower bound is proportional to the supremum of u. An analogous bound is obtained for a related Green's function.     

One-step t-fault diagnosis for hypermesh optical interconnection multiprocessor systems

To maintain high reliability and availability, system-level diagnosis should be considered for the multiprocessor systems. The self-diagnosis problem of hypermesh, emerging potential optical interconnection networks for multiprocessor systems, is solved in this paper. We derive that the precise one-step diagnosability of kⁿ-hypermesh is n(k − 1). Based on the principle of cycle decomposition, a one-step t-fault diagnosis algorithm for kⁿ-hypermesh which runs in O(kⁿn(k − 1)) time also is described.     

Two-step compensation of nonlinear systems

We study the problem of stabilization of nonlinear plants. We show that given a nonlinear plant P, if there exists a (nonlinear) compensator F, possibly unstable, which stabilizes P, then, with P₁: = P(I − F(− P))⁻¹, any C defined by C:= F + Q(I − P₁Q)⁻¹ for some finite-gain stable Q will stabilize P.     

Application of carbon ionic liquid electrode for the electrooxidative determination of catechol

A carbon ionic liquid electrode (CILE) was constructed using graphite powder mixed with N-butylpyridinium hexafluorophosphate (BPPF₆) in place of paraffin as the binder, which showed strong electrocatalytic activity to the direct oxidation of catechol. In pH 3.0 phosphate buffer solution (PBS) a pair of redox peaks appeared on the CILE with the anodic and the cathodic peak potential located at 387 and 330 mV (vs. SCE), respectively. The electrochemical behaviors of catechol on the CILE were carefully investigated, and the electrochemical parameters were calculated with the results of the electrode reaction standard rate constant k_s as 1.27 s⁻¹, the charge-transfer coefficient α as 0.58 and the electron transferred number n as 2. Under the selected conditions, the anodic peak current increased linearly with the catechol concentration over the range from 1.0 × 10⁻⁶ to 8.0 × 10⁻⁴ mol L⁻¹ by cyclic voltammetry at the scan rate of 100 mV s⁻¹. The detection limit was calculated as 6.0 × 10⁻⁷ mol L⁻¹ (3σ). The CILE showed good ability to separate the electrochemical responses of catechol and ascorbic acid (AA) with the anodic peak potential separation as 252 mV (vs. SCE). The proposed method was further applied to the synthetic samples determination with satisfactory results.     

A necessary condition for quantitative exponential stability of linear state-space systems

For an arbitrary n×n constant matrix A the two following facts are well known:

• (1/n)Re(traceA)−max_j=1,…,nRe λ_j(A)0;

• If U is a unitary matrix, one can always find a skew-Hermitian matrix A so that U=e^A.

In this note we present the extension of these two facts to the context of linear time-varying dynamical systemsAs a by-product, this result suggests that, the notion of “slowly varying state-space systems”, commonly used in literature, is mathematically not natural to the problem of exponential stability.