Gmd Forschungszentrum

Publications of Gmd Forschungszentrum

• Agents, a Broker, and Lies

  Authors: Peter Fankhauser, Thomas Tesch, Gmd Forschungszentrum, Informationstechnik Gmbh

  09/2000;

  Virtual enterprises need reliable and efficient cooperation mechanisms to carry out transactions between autonomous agents with conflicting interests. Available cooperation mechanisms either useVirtual enterprises need reliable and efficient cooperation mechanisms to carry out transactions between autonomous agents with conflicting interests. Available cooperation mechanisms either use bilateral multi-step negotiation or auctioning. Negotiations encourage agents to reason about the interests of their opponents. Thus, negotiations suffer from counterspeculations. Auctions apply to asymmetric trading only; they either favor the auctioneer or the bidders. Both mechanisms do not promote agents to tell the truth. Therefore, we propose to use a trustbroker to mediate between the agents. We introduce three symmetric, negotiation free one-step protocols to carry out a sequence of decisions for agents with possibly conflicting interests. The protocols achieve substantially better overall benefit than random or hostile selection, and they avoid lies. We analyze the protocols with respect to informed vs. uninformed lies, and with respect to beneficial vs. malevolent lies, and show that ...

• Resource Prediction and Admission Control for Interactive Video Browsing Scenarios Using Application Semantics

  Authors: Karl Aberer, Silvia Hollfelder, Gmd Forschungszentrum, Informationstechnik Gmbh

  09/2000;

  Highly interactive multimedia applications, like browsing in video databases, generate strongly varying loads on the media server during the presentation of media data. Existing admission controlHighly interactive multimedia applications, like browsing in video databases, generate strongly varying loads on the media server during the presentation of media data. Existing admission control approaches for limiting the number of concurrent users and thus guaranteeing acceptable service quality are only suited for applications with uniform load characteristics like video-ondemand. We propose an approach to admission control that is based on the stochastical model of Continuous Time Markov Chains, which allows to describe the different presentation states occurring in the interactive access to the multimedia database. The model is derived from semantic information on the forthcoming browsing session, in particular, it considers the relevance of the videos to the user. In this way a more precise prediction on resource usage can be given for achieving the two goals of Quality of Service (QoS) and good server utilization. The admission control mechanism is part of a multimedia database...

• Designing for Semantic Access: A Video Browsing System

  Authors: Silvia Hollfelder, Andr E Everts, Ulrich Thiel, Gmd Forschungszentrum, Informationstechnik Gmbh

  09/2000;

  Digital Library applications based on huge amounts of digital video data require efficient browsing and searching mechanisms for the extraction of relevant information. To avoid information overload,Digital Library applications based on huge amounts of digital video data require efficient browsing and searching mechanisms for the extraction of relevant information. To avoid information overload, the browsing system needs to preselect shots of interest from the database in a user-adequate manner. Additionally, it should support continuous presentation of time-dependent media. In this paper, a browsing system architecture is proposed that offers conceptual, content-based access to videos, and content-based buffering and admission control. It consists of the following components. The core of our architecture is a retrieval engine which calculates relevance values for the results of a conceptual query by feature aggregation on video shot granularity. An admission control mechanism supports highly interactive browsing scenarios and uses application semantics for the admission of new client sessions. We extended a commercial DBMS system for continuous media transport and presentation. A...

• Stochastic Resource Prediction and Admission for Interactive Sessions on Multimedia Servers

  Authors: Matthias Friedrich, Silvia Hollfelder, Karl Aberer, Gmd Forschungszentrum, Informationstechnik Gmbh

  09/2000;

  In multimedia systems, session-based admission control strategies are needed to guarantee QoS for a client session as a whole. This requires the prediction of irregular and varying resource demandsIn multimedia systems, session-based admission control strategies are needed to guarantee QoS for a client session as a whole. This requires the prediction of irregular and varying resource demands which result from interactive user behavior. The admission control scheme we present in this paper models the user behavior as Continuous Time Markov Chains (CTMCs). By performing transient analysis on the CTMC, the model is able to predict varying resource demands. This allows the detection of possible overload situations for future time intervals of specific length. Evaluation studies show that our approach obtains a good ratio of Quality of Service and, at the same time, a high server utilization. Keywords: Admission Control, Interactive Multimedia Applications, Continuous Time Markov Chains, Multimedia DBMS Zusammenfassung Um fur multimediale Sitzungen Qualitatsgarantien geben zu konnen, mu sich deren Zulassung auf die gesamte Sitzungsdauer beziehen. Bei stark interaktivem Nutzerverhal...

• An Approach For Managing Highly Configurable Operating Systems

  Authors: Danilo Beuche, Gmd Forschungszentrum, Informationstechnik Gmbh

  01/1999;

  Introduction Building a small and well adapted operating system for a particular problem class is one thing, building the same for a wide range of problems another. Embedded systems e.g. are becomingIntroduction Building a small and well adapted operating system for a particular problem class is one thing, building the same for a wide range of problems another. Embedded systems e.g. are becoming more and more complex, and the assembly style programming which is quite usual in this context does not allow a quick adaption of old sources to a new problem which is only slightly different. Similar problems arise in massively parallel systems. The operating system should support every application optimal i.e. have no feature that is not used by the running application. Different application have different needs, one uses only a single task, other are doing full multitasking. All these operating systems have so much in common, that it pays off to share those common parts. Both kinds of adaption are solvable through use of object-oriented designs. But object-orientation is often said to bear too much overhead to be practical for systems with very limited resources. We claim that

• Independent Component Analysis of Non-invasively Recorded Cortical Magnetic DC-fields in Humans

  Authors: Andreas Ziehe, Guido Nolte, Bruno-Marcel Mackert, Gabriel Curio, Gmd Forschungszentrum, Informationstechnik Gmbh, Schlo Birlinghoven

  01/1999;

  Artifacts in magnetoneurography (MNG) data due to endogenous biological noise sources, e.g. heart signal, can be four orders of magnitude higher than the signal of interest. Therefore it is importantArtifacts in magnetoneurography (MNG) data due to endogenous biological noise sources, e.g. heart signal, can be four orders of magnitude higher than the signal of interest. Therefore it is important to establish effective artifact reduction methods. We propose a blind source separation algorithm using only second order temporal correlations for cleaning bio-magnetic measurements of evoked responses in the peripheral nervous system. The algorithm showed its efficiency by eliminating disturbances originating from biological and technical noise sources and successfully extracting the signal of interest. This yields a significant improvement of the neuro-magnetic source analysis. Keywords Biomedical data processing, Artifact reduction, Biomagnetism, Magnetoneurography, Blind source separation, Independent component analysis. Acknowledgment We acknowledge valuable discussions with members of the Biophysics Group at PhysikalischTechnische Bundesanstalt (PTB) Berlin. All MNG studies were p...

• Experimental Symbolic Analysis of Net Systems

  Authors: Hartmann J. Genrich, Gmd Forschungszentrum, Informationstechnik Gmbh

  07/1998;

  this paper all quantities have values that are combinations of some pieces. These pieces can be viewed as di#erent units for expressing a value of the quantity -- like the various coins and notes ofthis paper all quantities have values that are combinations of some pieces. These pieces can be viewed as di#erent units for expressing a value of the quantity -- like the various coins and notes of a currency. Mere numbers themselves as values then are combinations of the unit ().

• Structure Oriented Case Retrieval

  Authors: Friedrich Gebhardt, Gmd Forschungszentrum, Informationstechnik Gmbh

  11/1997;

  While case similarity is mostly based on comparison of case attributes, many other possibilities exist. A variety of approaches found in the literature has been evaluated here. The guiding issues inWhile case similarity is mostly based on comparison of case attributes, many other possibilities exist. A variety of approaches found in the literature has been evaluated here. The guiding issues in this survey are the case structure and the similarity function. Important types of structure-based case retrieval procedures concern geometrical relations, graph algorithms, semantic nets and exploitation of deep (specific) domain models. 1 Case similarities Most case retrieval systems base the search on the comparison of descriptors or of values for attributes. This procedure is so widespread that even textbooks and survey articles ignore other possibilities. Yet there are many other ways to define similarity between cases and to organize retrieval accordingly. The common ground for these non-standard procedures is to exploit structures in the cases that go beyond the simple enumeration of some attributes. But apart from this commonality there exists a large variety in the case retrieval ...

• Microcontroller Selection Based on Retargetable Performance Prediction

  Authors: Paul G. Plöger, Gmd Forschungszentrum, Informationstechnik Gmbh, Schlo Birlinghoven, Heinrich Theodor Vierhaus, Enburgische Technische Universität Cottbus

  11/1997;

  In embedded system design the selection of an adequate controller is an decisive step. The demands of the application must be characterized and should match the candidate processor. We base thisIn embedded system design the selection of an adequate controller is an decisive step. The demands of the application must be characterized and should match the candidate processor. We base this selection on performance predictions and instruction set modeling. In this article we describe a way to automatically characterize controllers together with their respective compilers on the basis of a generic assembler. A generator produces test benches to measure and tabulate instruction timings on different CPUs. These are then used to predict the runtimes of the whole application running on the target hardware. Results for the table generation process, predictions and comparisons of different choices are given at hand of typical examples. 1 Introduction Especially in small and medium-sized companies a common methodology for the design of embedded systems proceeds as follows. Take a well-established development system and an existing system realization, change some necessary details and ch...

• A Group-based Authorization Model for Computer-Supported Cooperative Work

  Authors: Klaas Sikkel, Gmd Forschungszentrum, Informationstechnik Gmbh

  05/1997;

  Requirements for access control in CSCW systems have often been stated, but groupware in use today does not meet most of these requirements. There are practical reasons for this, but one of theRequirements for access control in CSCW systems have often been stated, but groupware in use today does not meet most of these requirements. There are practical reasons for this, but one of the problems is the inherent complexity of sophisticated access control models. We propose a general authorization model that emphasizes conceptual simplicity. Several extensions to the basic model address well-known issues in access control, notably negative rights and delegation. Implementation of the authorization model in the Basic Support for Cooperative Work Shared Workspace system is envisaged. Zusammenfassung In der Fachliteratur werden oft Anforderungen an der Zugriffskontrolle bei computergest utzte Zusammenarbeit aufgelistet. Die in der Praxis verwendeten Systeme kommen diesen Anforderungen jedoch nicht nach. Teilweise gibt es praktische Grunde dafur, es liegt aber auch an der inharente Komplexitat entprechender Zugriffsrechtsmodelle. Wir schlagen ein allgemeines Modell fur Zugriffsbere...

• The Synchronous Approach to Designing Reactive Systems

  Authors: Olivier Maffeis, Matthew Morley, Gmd Forschungszentrum, Informationstechnik Gmbh

  03/1997;

  Synchronous programming is available through several formally defined languages having very different characteristics: Esterel is an imperative language while Lustre and Signal are declarative inSynchronous programming is available through several formally defined languages having very different characteristics: Esterel is an imperative language while Lustre and Signal are declarative in style; Statecharts and Argos are graphical and allow one to program by constructing hierarchical automata. Our motivation for taking the synchronous design paradigm further, integrating imperative, declarative (or dataflow), and graphical programming styles, is that real systems typically have components that match each of these profiles. In this paper we motivate our interest in mixed language programming of embedded software around a number of examples, and sketch the semantical foundation of the Synchronie toolset which ensures a coherent computational model. This toolset supports a design trajectory that incorporates rapid prototyping and systematic testing for early design validation, an object oriented development methodology for long term software management, and formal ver...

• Methods and Systems for Case Retrieval . . .

  Authors: Druck Gmd, Forschungszentrum Informationstechnik Gmbh, Sankt Augustin, Friedrich Gebhardt, Gmd Forschungszentrum, Informationstechnik Gmbh

  01/1996;

  ion and Refinement in an Integrated System, a general planning system (e. g. for a lathe). The Fabel system ODM (see Section 3) can also be classified into this section; here the hierarchy levelsion and Refinement in an Integrated System, a general planning system (e. g. for a lathe). The Fabel system ODM (see Section 3) can also be classified into this section; here the hierarchy levels (di#erent resolutions) are artificially introduced for reducing search time. Chapter 8 Complementary issues 8.1 Loosely related work Some other case retrieval systems display interesting ideas but are either too specialized or too scarcely documented for inclusion in our classification. A somewhat unusual application is the focus of CAESAR: a library of 30 mathematical subroutines (linear algebra) [ Fouque and Matwin, 1993 ] . The query is stated in a Prologlike language; retrieval uses deductions and inductions on terms of a thesaurus. The exceptional feature of CADET, an attribute-based system, is a bunch of transformations using domain knowledge that can be applied to the query [ Sycara et al., 1992 ] . It is used for engineering design. Forest fires are the subject of another system [ ...

• Dynamic Priorities as a Basis for Control in Non-Sequential Systems

  Authors: Druck Gmd, Forschungszentrum Informationstechnik Gmbh, Sankt Augustin, Friedrich Gebhardt, Lauter Blabla, Gmd Forschungszentrum, Informationstechnik Gmbh, Forschungsbereich Kunstliche Intelligenz, Fakultat Informatik

  10/1995;

  this article. The missing pieces await experiences with modelling extended examples and with implementing the language. These missing bits are mentioned in Chapters 4 and 5. There is no urgent needthis article. The missing pieces await experiences with modelling extended examples and with implementing the language. These missing bits are mentioned in Chapters 4 and 5. There is no urgent need for fixing the strategy layer language; so this will also wait until more experience is gained. 6.2 Example problems

• MoMo: Language and Case Studies

  Authors: Druck Gmd, Forschungszentrum Informationstechnik Gmbh, Sankt Augustin, Friedrich Gebhardt, Gmd Forschungszentrum, Informationstechnik Gmbh

  08/1995;

  This report serves two purposes. Firstly, it documents the syntax and semantics of the modelling language MoMo: the complete specification of the task layer; changes against previous definitions toThis report serves two purposes. Firstly, it documents the syntax and semantics of the modelling language MoMo: the complete specification of the task layer; changes against previous definitions to domain and inference layers and to views connecting both; essential components of the strategy layer. Secondly, it discusses three detailed case studies to show the applicability of MoMo. This research was supported by the Federal Ministry of Education, Science, Research and Technology (BMBF) within the joint project FABEL under contract no. 01IW104. Project partners in FABEL are GMD -- German National Research Center for Information Technology, Sankt Augustin, BSR Consulting GmbH, Munchen, Technical University of Dresden, HTWK Leipzig, University of Freiburg, and University of Karlsruhe. BSR Consulting GmbH Wirtstraße 38 81539 Munchen Ansprechpartner: Dr. Brigitte Bartsch-Sporl Fax: 089/695158 Tel.: 089/695545 E-mail: brigitte@bsr-consulting.de GMD -- Forschungszentrum Informationstechnik GmbH Institut fur Angewandte Informationstechnik Forschungsbereich Kunstliche Intelligenz 53754 Sankt Augustin Ansprechpartner: Dr. Angi Voß Fax: 02241/14-2384 Tel.: 02241/14-2726 E-mail: angi.voss@gmd.de Technische Universitat Dresden Fakultat Informatik Institut Kunstliche Intelligenz 01062 Dresden Ansprechpartner: Doz. Dr. Uwe Petersohn Fax: 0351/4575-335 Tel.: 0351/4575-431 E-mail: peterson@freia.inf.tu-dresden.de HTWK Leipzig (FH) FB Informatik, Mathematik u. Naturwissenschaften Universitarer Studiengang Informatik Postfach 66 04251 Leipzig Ansprechpartner: Prof. Dr. Klaus P. Jantke Fax: 0341/328028 Tel.: 0341/3928-426 E-mail: jantke@informatik.th-leipzig.de Universitat Freiburg Institut fur Informatik und Gesellschaft Abt. Kognitionswissenschaft Friedrichstr. 50 79098 Freiburg Ansprec...