Showing papers by "Johannes Kepler University of Linz published in 1995"

A convergence analysis of the Landweber iteration for nonlinear ill-posed problems

Abstract: In this paper we prove that the Landweber iteration is a stable method for solving nonlinear ill-posed problems. For perturbed data with noise level \(\delta \) we propose a stopping rule that yields the convergence rate\(O (\delta ^{1/2}\) ) under appropriate conditions. We illustrate these conditions for a few examples.

How the sperm dominates the ovum — objectification by metaphor in the social representation of conception

Abstract: This investigation is about the use of metaphors in the everyday understanding of conception. It is argued that the analysis of the relationship between source and target domain in a metaphor used as an objectification device can help explain how social representations are acquired collectively and individually. We expect that in popular knowledge of conception the central metaphors and images relate to the subjects' everyday experience as social actors and sexual beings because social and sexual experience is pervasive and well-understood. In an experimental questionnaire study involving 169 subjects, it is shown as hypothesized that subjects prefer to compare the process of fertilization, i.e. the role and behaviour of sperms and ovum, with sexual and sex-role behaviour where the role of men is projected upon sperms and the role of women upon the ovum. This implies that sperms are seen not only as more active, but also as harder, stronger, and more dominant than the ovum. These effects are stronger, the more subjects personally subscribe to a more conservative sex-role orientation. In the discussion it is suggested that it is necessary to analyse the two intricately linked levels of objectification, the cognitive process of selecting specific images and the social process of the diffusion of popular knowledge, if we want to understand how common sense works.

DEVS Framework for Modelling, Simulation, Analysis, and Design of Hybrid Systems

Abstract: We make the case that Discrete Event System Specification (DEVS) is a universal formalism for discrete event dynamical systems (DEDS). DEVS offers an expressive framework for modelling, design, analysis and simulation of autonomous and hybrid systems. We review some known features of DEVS and its extensions. We then focus on the use of DEVS to formulate and synthesize supervisory level controllers.

Initialization of neural networks by means of decision trees

Abstract: The performance of neural networks is known to be sensitive to the initial weight setting and architecture (the number of hidden layers and neurons in these layers). This shortcoming can be alleviated if some approximation of the target concept in terms of a logical description is available. The paper reports a successful attempt to initialize neural networks using decision-tree generators. The TBNN (tree-based neural net) system compares very favourably with other learners in terms of classification accuracy for unseen data, and it is also computationally less demanding than the back propagation algorithm applied to a randomly initialized multilayer perceptron. The behavior of the system is first studied for specially designed artificial data. Then, its performance is demonstrated by a real-world application.

Nuclear magnetic resonance evidence for a new microstructure in ethene-cyclopentene copolymers

Abstract: We have analyzed the microstructure of a series of copolymers of ethene and cyclopentene prepared with a novel ferrocene-substituted catalyst using one- and two-dimensional 13 C- and 1 H-NMR methods. Up to 64% of the cyclopentene is incorporated as 1,3-cyclopentylene units, which has not been observed previously in copolymers. The subspectra were derived based on a triad model. The polymers appeared to be atactic with mainly cis-incorporated cyclopentylene units. No ring opening could be detected.

Social representations, group affiliation, and projection: Knowing the limits of validity

Abstract: It is argued that social representations and similar constructs can only be conceived of as mental structures containing meta-information about the group, within which the representation was formed and where it is part of social identity. The most important information will be to know the limits of its validity, that is, to know to which social group a representation pertains. Experiments on social projection have shown that people tend to project their opinions onto others, if they are perceived as being similar in background values. Such projection cannot be expected with idiosyncratic attitudes and beliefs. Consequently it is hypothesized that idiosyncratic and private attitudes, opinions, etc., are not attributed to specific social groups, whereas knowledge pertaining to social representations the subject subscribes to is projected onto the ingroup and less on the outgroup. This effect is expected to be independent of the relative number of people holding this opinion. This is what was found in a quasi-experimental questionnaire study. Implications of the findings are discussed with regard to the definition and criteria of social representations and to the structure of the theory

Generating multiple levels of detail from polygonal geometry models

Abstract: This paper presents a new method for solving the following problem: Given a polygonal model of some geometric object generate several more and more approximative representations of this object containing less and less polygons. The idea behind the method is that small detail in the model is represented by many spatially close points. A hierarchical clustering algorithm is used to generate a hierarchy of clusters from the vertices of the object’s polygons. The coarser the approximation the more points are found to lie within one cluster of points. Each cluster is replaced by one representative point and polygons are reconstructed from these points. A static detail elision algorithm was implemented to prove the practicability of the method. This paper shows examples of approximations generated from different geometry models, pictures of scenes rendered by a detail elision algorithm and timings of the method at work.

Finite Dimensional Generalized Baker Dynamical Systems for Cryptographic Applications

Abstract: In the past there have been several attempts to apply the field of deterministic chaos to cryptography. In this paper we propose for cryptographic applications to use maps as state transition function of a discrete dynamical system which leads to deterministic chaos. More specifically, we make use of generalized versions of the well-known baker transform, which are discrete and finite. Since they relate by group-theoretic representation to Bernoulli-shifts, we call them BernouIli permutations. The iteration of Bernoulli permutations on a set of data realizes a repeated "stretching" and "compressing" which has been compared by the "rolling" and "folding" in the work of a baker by mixing a dough. The knowledge of the importance of such operations for cryptography goes back to the fundamental paper of Claude Shannon ([8]), a fact which has been pointed out earlier in a paper by N.J.A. Sloane ([9]). The results reported in this paper are based mainly on the PhD-thesis of the second author ([6]). A Technical Report of the first author formed the starting basis for it. This paper extends the results which have been reported in earlier papers ([10],[11]).

Iv–vi quantum wells for infrared lasers

Abstract: Of central importance for mid-infrared diode lasers are their actual cryogenic operation temperatures. Quantum structures offer the potential for threshold current reduction and hence an operation temperature increase. The present experimental state of quantum structure diode lasers is reviewed. The relevant laser properties of the most promising material system, the IV–VI's are treated theoretically, and experimental results for IV–VI quantum well and superlattices samples are discussed.

Elastic strains in GaAs/AlAs quantum dots studied by high-resolution x-ray diffraction.

Abstract: We have studied a GaAs/AlAs periodic quantum dot array (fabricated by electron beam lithography and reactive ion etching) using high-resolution x-ray reciprocal space mapping around the (004) and (1\ifmmode\bar\else\textasciimacron\fi{} 1\ifmmode\bar\else\textasciimacron\fi{}3) reciprocal lattice points. Both the coherently and the diffusely scattered x-ray intensities were analyzed by performing two-dimensional model calculations and comparing them with the measured reciprocal space maps of the diffracted intensity. From the distribution of the diffracted intensities we deduced the average strain status in the dots. From the numerical simulations it is evident that random elastic strain fields are present, which extend through almost the whole volume of the quantum dot. The simulations of the x-ray measurements revealed that the crystalline part of the dots is considerably smaller as scanning electron micrographs would indicate, namely, 50 nm instead of 65 nm, respectively.

Simulation of NC machining based on the dexel model: A critical analysis

Abstract: Simulation systems for NC machining that also provide an analysis and verification component are a powerful tool to detect errors in NC codes off-line, ie before the NC program is put onto the machine and before damage can happen The modelling scheme used to represent the workpiece and its development during machining has a great influence on the behaviour and the efficiency of the whole system Without a well-chosen modelling scheme a realistic simulation and, especially, analysis and verification of NC programs are not possible Recently, at RISC-Linz, an NC simulation system, NC-SAVE, has been developed that is based on the dexel model The dexel model is an instance of spatial enumeration techniques, so it is an approximation technique, which allows the user to specify the accuracy he needs Practical experience with the system has shown that the dexel model allows an efficient dynamic update of the workpiece and provides many facilities for analysis operations In this paper, we study the accuracy of this modelling scheme More precisely, we give formulae that state, depending on the curvature of the objects to be modelled, which resolution to take for the model to achieve a certain (pre-specified) accuracy We also consider the influence of Boolean operations on the accuracy of models The question of the interdependence between accuracy and resolution of the model, and also the influence of various operations on the model is quite important, in order to choose the resolution of the model in such a way that tolerances are kept It turns out that the encoding of an object has to be done quite carefully in order to guarantee a certain accuracy of the model and to ensure that analysis operations and comparisons of models work properly The insights from the theoretical analysis of the dexel model are interpreted in terms of their influence on practical applications of the system NC-SAVE A much better understanding of the importance of good choices for the resolution of the model and the number and the positions of the viewpoints is provided It is not the intention of this paper to demonstrate the usefulness and the advantages of the system NC-SAVE It is rather the goal of this paper to give a careful analysis of the accuracy of the modelling scheme used within NC-SAVE Such an analysis, together with recommendations for the use of the system, improve the results of the simulation

Adapting to drift in continuous domains (Extended abstract)

Abstract: The experiments demonstrate that FRANN compares favourably with FLORA4 in the presence of concept drift. Learning is possible from examples described by symbolic as well as by numeric attributes, and because of its representation formalism (RBF networks, which realize a kind of prototype weighting scheme) FRANN is particularly effective in capturing concepts with nonlinear boundaries.

Automated support for software development with frameworks

Abstract: This document presents some of the results of an industrial research project on automation of software development. The project's objective is to improve productivity and quality of software development. We see software development based on frameworks and libraries of prefabricated components as a step in this direction. An adequate development style consists of two complementary activities: the creation of frameworks and new components for functionality not available and the composition and configuration of existing components.Just providing adequate frameworks and components does not necessarily yield automation and efficiency of software development. We developed the concept of relations between software components as a foundation for abstraction, reuse and automatic code generation for component interrelationship. Furthermore we suggest to supplement frameworks with an active cookbook consisting of active recipes which guide the software developer in the use of framework elements.In this paper our concept of using relations among software components is presented and the active cookbook is illustrated as a means for developer guidance. We created a prototype to demonstrate these concepts.

Designing a Framework by Stepwise Generalization

Abstract: Frameworks are class collections that capture the common aspects of applications in a certain problem domain. It is argued that conventional object-oriented design methods do not suit well to frameworks. As an alternative, we suggest a two-phase design method. The first phase is a stepwise generalization of a problem to its most general form. The second phase proceeds in the opposite direction, giving an implementation for every generalization level. The implementation of a certain level can be regarded as a framework for implementing the next more specific level. The main advantage of our method is that it enhances the generality (and therefore reusability) of the framework architecture, and explains the design decisions in the simplest possible context. As a case study, we describe a framework supporting the analysis of formal languages.

Determination of chelating agents and metal chelates by capillary zone electrophoresis

Abstract: Aminopolycarboxylic acids such as diethylenetriaminepentaacetic acid (DTPA) are commonly used as chelating agents in many pulp and paper industries, particularly as scavengers of metal ions which catalyze the decomposition of hydrogen peroxide used as a bleaching agent. Concern for the effect of waste DTPA in the aquatic environment has led to a need for the development of methods to determine its levels in waste water. This paper describes the determination of free DTPA and several metal-DTPA complexes in water and waste water by capillary zone electrophoresis. The optimization of separation conditions included the selection of an appropriate carrier electrolyte composition (pH, organic solvents, ion-pairing reagents) and the systematic investigations of selective complexation of free DTPA as well as metal exchange reactions for metal-DTPA complexes in order to achieve selective and sensitive direct UV detection. The determination of DTPA in waste water from a paper mill was possible in the low ppm range.