The Kantorovich Metric in Computer Science: A Brief Survey

In contrast to its wealth of applications in mathematics, the Kantorovich metric started to be noticed in computer science only in recent years. We give a brief survey of its applications in probabilistic concurrency, image retrieval, data mining, and bioinformatics. This paper highlights the usefulness of the Kantorovich metric as a general mathematical tool for solving various kinds of problems in rather unrelated domains.     

Local boundary estimates for the Lagrange multiplier discretization of a Dirichlet boundary value problem with application to domain decomposition

Abstract We give an estimate on the error resulting from approximating the outer normal derivative of the solution of a second-order partial differential equation with the Lagrange multiplier obtained in using the Lagrange multiplier method for imposing the Dirichlet boundary conditions. We consider both the case of smooth domains and, in view of an application in the framework of domain decomposition, the case of polygonal domains. The estimate given, supported by numerical results, shows that the mesh in the interior of the (sub)domain can be chosen more coarsely than near the boundary, when only a good approximation of the outer normal derivative is needed, as in the case of the evaluation of the Steklov-Poincaré operator, or when solving with a Schur complement approach the linear system arising from the three-fields domain decomposition method. Keywords: Lagrange multipliers, error estimate, normal derivative, Steklov-Poincaré operator     

An efficient hybrid reduction method for time-delay systems using Hermite expansions

This paper presents an efficient model reduction method for time-delay systems in the time domain. We expand the systems under a Hermite polynomial basis and show that Hermite coefficients of the expansion are determined by a linear equation, thus can be calculated efficiently. Such linear relationship is well taken in the projection methods of model reduction, and reduced models are generated to preserve a desired number of Hermite coefficients in the time domain, in contrast to other existing techniques which aim at approximating the transfer function of time-delay systems in the frequency domain. We also exploit two-sided projections for time-delay systems, leading to a hybrid reduction method which generates reduced models sharing the nice properties both in the time and frequency domains. Two numerical examples illustrate the feasibility and effectiveness of the approach.     

Metric completion versus ideal completion

Complete partial orders have been used for a long time for defining semantics of programming languages. In the context of concurrency de Bakker and Zucker (1982) proposed a metric setting for handling concurrency, recursion and nontermination, which has proved to be very successful in many applications. Starting with a semantic domain D for ‘finite behaviour’ we investigate the relation between the ideal completion Idl(D) and the metric completion which are both suitable to model recursion and infinite behaviour. We also consider the properties of semantic operators.     

时序逻辑:线性时间对分支时间

1.引言 时序逻辑是一种模态逻辑,不同于经典逻辑,其真假值依赖于时间而变化。对于序的数学处理,并不是静态的,它包含很重要的动态特征。一个典型的程序执行模型就是状态列,不同的状态,程序的属     

Multi-domain learning by confidence-weighted parameter combination

State-of-the-art statistical NLP systems for a variety of tasks learn from labeled training data that is often domain specific. However, there may be multiple domains or sources of interest on which the system must perform. For example, a spam filtering system must give high quality predictions for many users, each of whom receives emails from different sources and may make slightly different decisions about what is or is not spam. Rather than learning separate models for each domain, we explore systems that learn across multiple domains. We develop a new multi-domain online learning framework based on parameter combination from multiple classifiers. Our algorithms draw from multi-task learning and domain adaptation to adapt multiple source domain classifiers to a new target domain, learn across multiple similar domains, and learn across a large number of disparate domains. We evaluate our algorithms on two popular NLP domain adaptation tasks: sentiment classification and spam filtering.     

Multi-level transaction model for semantic concurrency control in linear hash structures

In this paper, we present a version of the linear hash structure algorithm to increase concurrency using multi-level transaction model. We exploit the semantics of the linear hash operations at each level of transaction nesting to allow more concurrency. We implement each linear hash operation by a sequence of operations at lower level of abstraction. Each linear hash operation at leaf-level is a combination of search and read/write operations. We consider locks at both vertex (page) and key level (tuple) to further increase concurrency. As undo-based recovery is not possible with multi-level transactions, we use compensation-based undo to achieve atomicity. We have implemented our model using object-oriented technology and multithreading paradigm. In our implementation, linear hash operations such as find, insert, delete, split, and merge are implemented as methods and correspond to multi-level transactions.     

Coalgebraic semantics for timed processes

We give a coalgebraic formulation of timed processes and their operational semantics. We model time by a monoid called a “time domain”, and we model processes by “timed transition systems”, which amount to partial monoid actions of the time domain or, equivalently, coalgebras for an “evolution comonad” generated by the time domain. All our examples of time domains satisfy a partial closure property, yielding a distributive law of a monad for total monoid actions over the evolution comonad, and hence a distributive law of the evolution comonad over a dual comonad for total monoid actions. We show that the induced coalgebras are exactly timed transition systems with delay operators. We then integrate our coalgebraic formulation of time qua timed transition systems into Turi and Plotkin’s formulation of structural operational semantics in terms of distributive laws. We combine timing with action via the more general study of the combination of two arbitrary sorts of behaviour whose operational semantics may interact. We give a modular account of the operational semantics for a combination induced by that of each of its components. Our study necessitates the investigation of products of comonads. In particular, we characterise when a monad lifts to the category of coalgebras for a product comonad, providing constructions with which one can readily calculate.     

Multidimensional Timestamp Protocols for Concurrency Control

We propose multidimensional timestamp protocols for concurrency control in database systems where each transaction is assigned a timestamp vector containing multiple elements. The timestamp vectors for two transactions can be equal if timestamp elements are assigned the same values. The serializability order among the transactions is determined by a topological sort of the corresponding timestamp vectors. The timestamp in our protocols is assigned dynamically and is not just based on the starting/finishing time as in conservative and optimistic timestamp methods. The concurrency control can be enforced based on more precise dependency information derived dynamically from the operations of the transactions. Several classes of logs have been identified based on the degree of concurrency or the number of logs accepted by a concurrency controller. The class recognized by our protocols is within D-serializable (DSR), and is different from all previously known classes such as two phase locking (2PL), strictly serializable (SSR), timestamp ordering (TO), which have been defined in literature. The protocols have been analyzed to study the complexity of recognition of logs. We briefly discuss the implementation of the concurrency control algorithm for the new class, and give a timestamp vector processing mechanism. The extension of the protocols for nested transaction and distributed database models has also been included.     

Improving concurrency in temporal plans

There is an increasing interest in solving temporal planning problems. Identification and propagation of mutual exclusion relations between actions can significantly enhance the efficiency of a planner. Current definitions of mutually exclusive actions severely restrict their concurrency. In this paper, we report on thirteen groups of permanently mutually exclusive PDDL 2.1, Level 3 actions. We report on sixteen types of potentially-conflicting interactions between two actions where concurrency may be maximized by adjusting starting time of one of the two actions. We discuss several examples where actions can overlap despite conflicting preconditions and/or effects. The processes executing these actions are mostly independent. We report on a new domain-rewriting technique called “baiting” in order to improve the concurrency in temporal plans. Baiting actions lure a temporal planner into improving concurrency. The technique involves splitting user-identified operators. We report on three types of baiting (standard, double and nested) and show their suitability for various types of action interactions. Baiting requires minimal modification to the planning code. Baiting does not increase the branching in search trees. Baiting does not affect the soundness and completeness of a temporal planner. Our empirical evaluation shows that the makespans of plans generated by efficient planner Sapa with baited domain are significantly lower than makespans of plans generated without baiting.     

On linear logic planning and concurrency

We present an approach to linear logic planning where an explicit correspondence between partial order plans and multiplicative exponential linear logic proofs is established. This is performed by extracting partial order plans from sound and complete encodings of planning problems in multiplicative exponential linear logic. These partial order plans exhibit a non-interleaving behavioural concurrency semantics, i.e., labelled event structures. Relying on this fact, we argue that this work is a crucial step for establishing a common language for concurrency and planning that will allow to carry techniques and methods between these two fields.     

Uncertainty principle for linear canonical transform using matrix decomposition of absolute spread matrix

We define a matrix, the absolute spread matrix (ASM), whose positive definite property is independent of signals. Based on the relation between the ambiguity function and the linear canonical power spectra, we give a new perspective on a key relation associated with spreads in time, frequency, time-frequency and linear canonical transform (LCT) domains, from which a new LCT uncertainty principle for complex signals using the ASM's rotation orthogonal decomposition is derived. The proposed lower bound of uncertainty product in two LCT domains is tighter than the latest one given by the author. We then obtain the sufficient and necessary condition that gives rise to this stronger result truly, and define a quantitative index to analyse the difference with the existing bound. We also reduce the uncertainty inequality to time and LCT domains. Furthermore, we present the example and simulation results that validate the previous theoretical analyses, and finally we demonstrate the effectiveness of the new proposal through discussing its applications in signal processing and optics.     

时延Petri网分布式模拟的先行值研究

先行值计算是提高时延Petri网并行模拟性能的一个好的方法。给出了时延Petri网的先行值计算的四种基本结构,对于存在循环的复杂的Petri网结构给出了预测图算法,通过预测图,能够很容易求出静态和动态先行值,在并行模拟中利用先行值可以分析出存在并发和阻塞的结构,从而为网分块在并行机的结点上运行奠定了基础。     

An enhanced concurrency control scheme for multidimensional index structures

We propose an enhanced concurrency control algorithm that maximizes the concurrency of multidimensional index structures. The factors that deteriorate the concurrency of index structures are node splits and minimum bounding region (MBR) updates in multidimensional index structures. The properties of our concurrency control algorithm are as follows: First, to increase the concurrency by avoiding lock coupling during MBR updates, we propose the PLC (partial lock coupling) technique. Second, a new MBR update method is proposed. It allows searchers to access nodes where MBR updates are being performed. Finally, our algorithm holds exclusive latches not during whole split time but only during physical node split time that occupies the small part of a whole split process. For performance evaluation, we implement the proposed concurrency control algorithm and one of the existing link technique-based algorithms on MIDAS-III that is a storage system of a BADA-IV DBMS. We show through various experiments that our proposed algorithm outperforms the existing algorithm in terms of throughput and response time. Also, we propose a recovery protocol for our proposed concurrency control algorithm. The recovery protocol is designed to assure high concurrency and fast recovery.     

An environment for formal verification based on symbolic computations

We present an environment for formally verifying hardware, based on symbolic computations. This includes a new concurrency model, called the combinational/sequential or C/S concurrency model which has close ties to hardware. We allow fairness constraints and describe methods for specifying them and for formally verifying in their presence. Properties are specified by either CTL formulae or edge-Rabin automata. We give algorithms, in the presence of fairness constraints, for model checking CTL or for checking that the language of our system is contained in the language of a property automation. Finally, techniques are given for hierarchical verification and for detecting errors quickly (early failure detection).     

并行数据库实时多版本并发控制协议性能研究

随着对实时数据库事务处理性能要求的不断提高 ,将并行数据库和实时数据库相结合的并行实时数据库将成为新的发展方向 .然而通常的实时多版本并发控制协议不适合并行数据库的无共享结构 .比较了几种并发控制协议在无共享结构下的性能 ,该文提出了一种并行实时多版本并发控制协议 .它能有效地减少事务的重启 ,降低事务的错过率 .在长短事务混合的情况下 ,它的性能比其它的并发控制协议更好 .同时 ,它具有更好的可扩展性 .作者利用仿真模型对该协议进行了性能测试 ,与其它协议进行了比较 ,并分析了实验结果 ,仿真实验结果表明该协议性能良好     

Bundle event structures: A revised cpo approach

Bundle event structures equipped with a partial order ? have been used to give a true concurrency denotational semantics for LOTOS. This model has also been extended by time and stochastic information. Unfortunately it fails to yield a complete partial order (cpo) as we illustrate by an example.We propose a subset of all bundle event structures such that it forms a cpo. This subset is closed under the usual operators on bundle event structures. And as a consequence these operators are continuous. Therefore, this subset can be used to give a denotational semantics of LOTOS.     

A self‐aligned multi‐domain liquid‐crystal display on polymer gratings in a vertically aligned configuration

Abstract— We report on a new method of fabricating a vertically aligned multi‐domain liquid‐crystal display (LCD) using surface‐relief gratings. A linear array of surface‐relief gratings was produced by using a photosensitive polymer material coated on glass substrates by the illumination of the UV light through a photomask. The LCD cell was assembled with two substrates with polymer gratings in such way that the grating vectors were orthogonal to each other. In this LCD configuration, the nematic molecules were reoriented by distortions of an external electric field at the grating surfaces to make four different domains. The LC cell with self‐aligned four domains shows excellent extinction in the off‐state and wide‐viewing characteristics in the on‐state.     

On the Linear Stability of the Fifth-Order WENO Discretization

We study the linear stability of the fifth-order Weighted Essentially Non-Oscillatory spatial discretization (WENO5) combined with explicit time stepping applied to the one-dimensional advection equation. We show that it is not necessary for the stability domain of the time integrator to include a part of the imaginary axis. In particular, we show that the combination of WENO5 with either the forward Euler method or a two-stage, second-order Runge–Kutta method is linearly stable provided very small time step-sizes are taken. We also consider fifth-order multistep time discretizations whose stability domains do not include the imaginary axis. These are found to be linearly stable with moderate time steps when combined with WENO5. In particular, the fifth-order extrapolated BDF scheme gave superior results in practice to high-order Runge–Kutta methods whose stability domain includes the imaginary axis. Numerical tests are presented which confirm the analysis.