Computer Aided Systems Theory - EUROCAST’99

The origin of Systems Theory lays in the kind of complex problems experienced by engineers (specifically in the field of communications and control) and scientists (in biology and ecology) in the mid of the 20th century. Then it became obvious, that the usual mathematical modeling concepts based on analysis (differential equations) and linear algebra (linear equations and matrices) were not any more appropriate. New mathematical methods in dealing with actual problems in formal modeling tasks were required.

The systematic use of what we now call Systems Theory in the description of biological systems, and more precisely, the nervous system, took off in the Forties although many of the basic ideas had been being managed in philosophic and scientific circles almost since the Ancient Greeks. From 1943 to 1945, a kind of synergetic process was started up, triggered as the result of three basic works. First, Norbert Wiener, Arthur Rosemblueth and Julian Bigelow’s study (1943) on the nature of teleological processes where the crucial idea was that what was relevant in a homeostatic process was the information return and not theenergy return via the feedback links. It is representative of the analytical approach.

The field of robotics is one of the most innovative in the last decade. Conventional industrial robots from the late 70’s are now only a tool on the production level. One of the oldest dreams of the robotic community – intelligent, mobile and humanoid robots – starts to become a reality because of the rapid development of “external” sensors.External sensors (e.g. visual, auditive, force-torque) offer intelligent robots the possibility to see, hear, speak, feel, smell like humans. Compared with conventional, unintelligent, industrial robots, intelligent robots can fulfill new, innovative tasks in new application areas.

The Real-Time Control System (RCS) Reference Model Architecture provides a well-defined strategy for development of software components for applications in robotics, automated manufacturing, and autonomous vehicles. ADLs are formally defined languages for specification of software system’s designs. In this report, we describe the results of an investigation into the use of an ADL to specify RCS software systems, and assess the potential value of ADLs as specification and development tools for RCS domain experts. The report also discusses potential influence of ADLs for commercial software development tools and component-based development.

Conceptual Design methods place some intermediate solution steps between problem definition and final technical solution. Functional Decomposition is such a method which aims at finding solution concepts in terms of functions first and, after that, technical realisations for the individual functions. The role of mathematical models in conceptual design is to provide some coarse quantitative assessment of certain solution concepts, and furthermore, to clarify the relationship between the Functional Requirements and the Design Parameters. Later on, in the more detailed design steps refined mathematical models are used. What is their relation to the early models used in conceptual design? The concept of perturbation analysis seems to provide a framework for defining and understanding these relationships.

In order to model complex systems it is indispensable to structure the modelings. The mechanism of abstraction as well as the usage of views is very helpful for structuring. In this paper I will discuss these aspects in the context of Petri systems in order to get fundamental definitions. It turns out, that with respect to many different definitions of Petri systems only a “generic definition” is possible. To get final definitions for concrete types of Petri systems it requires much further work.

In applications of formal methods, the problem of how to establish the correctness of the initial formalization step is an often underestimated aspect of the system design process. We propose a methodology based on the construction of a mathematical model which reflects the given system so closely that the correctness can be established by observation and experimentation (ground model). Complex technical systems are often heterogeneous, so that different system aspects are best modelled by means of different techniques. This leads to heterogeneous ground models. To achieve a consistent and coherent view of heterogeneous behavioural models, which is a prerequisite for any systematic analysis and validation, we introduce a common semantic framework (meta-model) based on the notion of Abstract State Machines. We exemplify our methodology by an industrial case study from automated manufacturing, the distributed control for a material flow system (MFS).

This paper deals with embedded systems architecture components called as application patterns, and with their employment for design reuse. The first part of this contribution introduces the concepts of application patterns and relates them to the well-known object-oriented design abstractions. Employing application patterns that demonstrate the concrete examples of reusability, the kernel of this contribution presents two case studies, which are based on real design projects: petrol pumping station dispenser controller and multiple lift control system. To reuse an architectural component whose implementation usually consists both of software and hardware, it means to reuse its formal specification. The paper deals with behavioral specifications employing state or timed-state sequences and with their closed-form descriptions by finite-state or timed automata. The contribution focuses on identification, creation, and initial classification of reusable application patterns while retrieval, adaptation, and storage reuse tasks with case-based reasoning support are treated briefly at the conclusion as an introductory information about launching research.

The article discusses the notion of state spaces of object-oriented Petri nets (OOPNs) associated to the tool called PNtalk and the role of identifiers of dynamically appearing and disappearing instances within these state spaces. Methods of working with identifiers based on sophisticated naming rules and mechanisms for abstracting names are described and compared. Some optimizations of state space generating algorithms for the context of OOPNs are mentioned, as well. Finally, some possibilities of specifying properties of systems to be checked over the state spaces of their OOPN-based models are discussed.

This paper addresses the possibility of representing and simulating Petri Nets using an Artificial Intelligence formalism for Reasoning about Actions and Change called Generalized Magnitudes. As a result, we obtain an alternative logical-based formalization for Petri Nets, allowing to apply common techniques in Reasoning about Actions and Change in the analysis of Petri Nets models. Futhermore, we provide the alternative of representing concurrency and synchronization in a Generalized Magnitudes system using a Petri Net.

The formal proof of properties of a system first requires the expression of the behavior of the system into a formal language.The scope of this paper is the simplification of the proof procedure of properties of systems which are represented by discrete time models (Finite State Machines or extensions) The formulas which are manipulated are decomposed and the global proof is reduced to the study of a small subset of elementary proofs. This method was obtained by re-using some work developed in the framework of the management of the Path Condition in Symbolic Simulation.

In this paper we derive exemplarily a parallel processor array for algorithms of commonly used tomographic reconstruction methods by using the tools of the design system DESA. The algorithms represent a group of computationally intensive image processing algorithms requiring high throughput and real-time processing.The design process is characterized by the consideration of hardware constraints and performance criteria. In particular, we determine one common parallel processor array for two different reconstruction techniques. Finally, the array is adapted to hardware constraints given by the target architecture which can be an application specific integrated circuit or a system of parallel digital signal processors.

The paper presents an approach to formalize a continuous-time system description and a suitable approach to elaborating intelligent computer aided design system. Corresponding principles and a structure scheme are proposed to implement the approach. The approach enables one to use both design rules and design algorithms by an effective manner in dependence on task requirements. The organization of calculations is considered under modeling continuous-time systems on the basis of using flexible simulating complex. Also, an application of the flexible simulating complex to design a universal model of a nuclear power plant is considered.

Arthur Koestler (1967, 1969, 1978) introduced into science the concepts of a “holon” and of a “holarchy”. A holarchy is according to A. Koestler a tree-like hierarchy where the nodes of the tree – the components of the hierarchy – are autonomous intelligent acting I/O systems.Koestler calls such components “holons”. Holarchies provide according to Koestler the appropriate conceptual framework for modeling and simulation of self organizing open hierarchical systems as they can be found in biology, medicine, sociology and management science. Especially he was interested to use holarchies for modeling the function of the human brain. In the past years artificial intelligence developed the topic of agent systems. Agents are intelligent acting systems which are able to perform certain tasks in an autonomous way. For the user agents are usually invisible. In case that agents are information processing systems their realization can be done in software (software agents). The elaboration of agent methodology led to the concept of a “multi-agent”, a network of cooperating agents (e.g. J. Ferber 1999). Agents have a conceptual similarity to the holons of A. Koestler. It seems to be natural to try to make use of the available agent technology to construct a network of systems which constitute a holarchy in the sense of A. Koestler. We call such system a Multi-Agent Holarchy.In this paper we try to make a first step towards such an investigation. Besides of the work of Koestler and the existing models for multi-agent systems we base our paper here on earlier work of this author (Pichler 1995, 1999).

In this paper we study the partition of a systems by a new theory denoted General System Logical Theory GSLT). The partition of a system in subsystems creates a conflict situation. In fact when we move from the system to the subsystems, new variables grow up. These variables are the elements that represent the interaction of one subsystem with the others. The interaction is computed by recursive equations that compute the interaction terms. So any subsystem can be isolated from the total system but we must join external variables or sources or inputs that represent the action of the other subsystems. The GSLT language is used to detect the conflict between the system and the subsystems and solve the conflict by the computation of the internal action among the subsystems. We can associate with any subsystem a global variable that changes in time only for the action of the other subsystems.

Within the last years the paradigm of intelligent software agents became an emerging topic in computer science research and development. Intelligent software agents are computational systems that populate complex dynamic environment, act and react there in an autonomous way in order realize a specific task based on a set of goals given to them. In multi agent systems (= MAS) intelligent software agents are grouped together and represent a network of problem solvers designed to achieve a common goal, which would be too large for a single centralized software agent. The member agents of such a multi agent system have a sufficient degree of decision-making autonomy and interact with other agents by explicit communication. Therefore the main fields of interest in multi agent based system design cover coordination, communication and negotiation based on uncertain data and knowledge of each member agent of the multi agent system. This article proposes s system theoretical approach towards the design and synthesis of a multi agent system for controlling a robotic agent.

The problem, which was addressed in this project, was the creeping loss of competence of the user when using conventional knowledge based systems. Therefore two main points were the aims of the project. At first it was concentrated on using the Multi-Agent architecture. The second aim was the development of design guidelines for competence promoting decision support systems. The promotion of competence was realised by offering different opinions about a problem to force decisions of the user. Two agents with different views on the diagnosis problem were implemented. Diagnosis strategies and data were analysed with regard to the different views and the integration into the system. At last methods to measure the effect on the competence of the user were developed and tested.

Intelligent robots are systems which process on-line a vast volume of information consisting of both the data and knowledge. The data gathered by sensors are processed making use of knowledge incorporated in the software modules. To explore both the data and knowledge requires to integrate both of them on appropriate levels. As different the data usually represent the same part of the world, it is obvious to speak about data fusion than integration. On the other hand, the software modules are usually highly specialized and complementary to each other. Solving the problems of relevant data fusion and systematic software integration seems to be a crutial aspect of the robot design tasks. The paper describes core ideas and principles used for software and system integration in the GLbot (the Gerstner Laboratory Robot) experimental platform, were the multi-agent approach has proved to be the very efficient one.

The research activities carried out in the framework of the proposed paper have been directed to the development of nonlinear control techniques, algorithms and architectures for the free motion of robots having rigid links, that provide the following global control performances: global asymptotic stability, precision for time varying trajectory tracking, robustness at external disturbances, nonlinear coupling between motion axes and uncertainties of the dynamic model of the manipulator with drives structure, for closed-loop robot systems with n ≥ 5 degrees of freedom. Thus, there is proposed a solution for nonlinear motion control based on adaptive control techniques. The control method that is proposed provides high dynamic performances for the ensemble manipulator – drive systems – internal transducers – controller – user interface, that is: high displacement speed and tracking precision on desired time varying trajectories, rejection of disturbances of the type: variable masses carried by the arm, modeling errors, measuring noise. In the framework of this paper, the authors designed and simulated a library of procedures, algorithms and software control modules to be further implemented in a generic parallel multiprocessing architecture, as a robot controller. In this respect, the structural design was directed towards VME – based bus oriented multimaster structures.

An approach to the mobile robot path planning among heterogeneous regions is presented. Proper weight factors are associated with the regions of different kind (rock, sand, grass) to characterize the amount of the difficulty required to travel through that particular region, as compared to traveling over a flat, smooth surface (corresponding to a minimum weight). The weighted regions are represented in the form of modified matrix quadtree. The technique of distance transform is extended to the weighted regions and applied to the robot path planning. Illustrative examples are included.

The article describes construction of a mathematical model of a robot The method is based on a matrix calculus. The advantage of the described method is in unified approach which allows systematic construction of mathematical model of mechanical part of a robot together with its electrical drives. Matrix engine of the MATLAB-SIMULINK then allows easy simulation of the complete robot. The article also discuss classical way of model construction and gives its comparison with the matrix approach.

In order to engineer business processes a model is needed that allows us to analysis and design of the dynamic properties of the processes. This paper shows a unified framework for data flow diagrams (DFD, for short) and Petri nets in modeling business processes. Introducing the concept of business transaction Petri net, we provide a firm mathematical basis for a comparison of a such Petri net to a business process modeled by a DFD. It will be turned out that the behavior of a business transaction Petri net is simulated by the corresponding business transaction system.The resulted framework makes it possible that the experience with DFD’s will be applicable to Petri nets in modeling business processes, and vice versa.

Market demands and internal needs forced companies to permanent adaptations. It exists a need for a systematic approach to the perpetual business process improvement and company growth. Reengineering of assembly is one of the possibilities, and involves activities on assembly technology, logistic, assembly system development and product redesign. For that reason, the adequate tools for modelling of the assembly process and system have been developed. Activities needed for reengineering process are part of such model. The reengineering process is finished and started again by evaluation phase of reengineering.

As we left the 20th century the industry has paid particular attention to the ecological impacts of new facilities, energy use, number of used materials and several other very important environmental issues. Now many impacts to design and recycle facilities as “life cycle” (planning, design, development, supply, manufacture, use, recycling) have not been systematically and comprehensively studied and analysed. Engineers, operators, professionals and other people who works in this area together maybe in different firms should understand the environmental and interactions of management, team behavior and methods because their everyday decisions carry substantial social implications.

This paper proposes a new approach to a rarely investigated topic: How can the emergence of new agents in social life be modeled. It proceeds in two steps. First a set of general simulation features needed to embed the notion of emergence are developed. Then it is shown how these challenges could be met by the use of a certain simulation environment called HE (Human Evolution). It turns out that not only specification and information structures used in HE are radically different from other contemporary approaches, but that the nature of results too leads researchers in a new – in my opinion highly prosperous – direction.

The paper presents the concept of an interactive team-building tool for simulating social grouping. It shows how decisions whom to assign which task can be supported from a game-theoretic perspective. Thus the group member are modeled as social entities following their preferences and being able to resist against work distributions they do not want. The features of ITBT are deduced from the decision situation of a team leader being in charge for the work distribution and are described in more detail in the paper.

In this article, we want to demonstrate that data acquisition and processing are not only technical and information problem, but that they have to fulfil some sociological aspects, too.The data acquisition and processing is the part of a cybernetic system. However, nowadays it is too often identified only with an information processing system itself. Such a contraction is not harmful. However, by perpetuating it we lose a fundamental sense of the activity – it becomes incomprehensible and ‘art for art’s sake’.

This paper studies the concurrent simulation of DEVS systems by using their encoding into a formalism for dynamic systems called Generalized Magnitudes (GM). When represented with GMs, the internal parallelism of a DEVS model is foregrounded and the concurrent inference techniques developed for the GMs formalism can be applied to speed up simulation. This approach is used for both atomic and coupled DEVS models, and does not require the modeler’s intervention. The internal structure of a typical DEVS control system is identified, so that the scheduling algorithms can be adapted for further efficiency improvement.

Motivated by a typical and well-known problem of neurobiological modeling, a parallel algorithm devised to simulate sample paths of stationary normal processes with rational spectral densities is implemented to evaluate first passage time probability densities for time-varying boundaries. After a self-contained outline of the original problem and of the involved computational framework, the results of numerous simulations are discussed and conclusions are drawn on the effect of a periodic boundary and a Butterworth-type covariance on determining quantitative and qualitative features of first passage time probability densities.

The training of the personnel plays a major role in the security and reliability of many industrial processes. In order to fulfil this objective it is necessary to develop applications that support distributed interactive simulation with an interface near the real world. This paper describes such type of application during the development of a training simulator prototype for a nuclear power plant. The simulator is based on a client/server architecture that allows the execution on different machines in a network and many users to participate in the same simulation. The interface was designed to support the interaction of the operators with the simulator through touch screens with high fidelity displays of the control room developed using the component technology. The main features of the simulator are the distributed execution of the models using inexpensive hardware and the flexibility of design and maintenance of the interface.

Starting from a continuous follow-the-leader model for individual vehicle motion a stochastic term is added in order to account for the elementary randomness of driving. In a suitably defined stationary state thus an equilibrium distribution of nearest-neighbour distances is spanned up which allows to understand the “fundamental diagram” of macroscopic traffic flow modelling as the ensemble average of the microscopic characteristic function. Estimates for the distribution width and a comparison with simulation results are given.

A model for the analysis of floating car data (FCD) in urban road networks is proposed. In contrast to freeway traffic, structural features such as traffic lights are shown to dominate the dynamics – in a quasistationary picture – of such a network. Principles for obtaining a segment travel time depending on the mean vehicle density are discussed. Some remarks for the determination of this state variable are given, also on several segments.

The influence of the hologram division into parts on the height of the readout signal was considered. It was found that from the two waves onto which the signal can be divorced: the geometrical and the boundary wave, the latter is responsible for the transfer of the optical signal.

After a short introduction to the concept of automation and computer based systems, chapter 2 presents the the concepts of Electronic Performance Support Systems (EPSS). Chapter 3 specifically discusses process-centered EPSS and their realization in software engineering and office automation. Different attitudes and requirements are pointed out. Chapter 4 argues that computer support makes it necessary to extend Scheer’s Y (well-known in the CIM-area) to accomodate additional granularity. This discussion is carried over into chapter 5 where we map the relevant processes of software engineering and office automation to the processes identified in the extended Y from chapter 4.

In the software development process ideas and objectives are evolved into semiformal and formal representations allowing later execution of the system on a computer. Knowledge about the system is represented in various forms (e.g., natural language, UML models, source code, etc.). The evolutionary change of knowledge representations is a challenging issue. In this paper we propose a framework applicable to the elicitation and evolution of requirements in the software engineering process. We discuss architectural considerations for supporting traceability between artifacts of different CASE tools and present a partial implementation of the approach.

In this paper we propose a new combinatorial optimization method for parts matching in a precision assembly system. This method combines selective assembly with micro machining so as to produce high precision assembled units. From the simulation results, it is presented that the system is effective in a small lot production, where higher assembly rates can not be attained with conventional methods, in such that a higher rate is obtained if micro machining is used alongside.

The assembly of complex microsystems consisting of several single components (i.e. hybrid microsystems) is a task which has to be solved to make mass production of microsystems possible. Therefore, it is necessary to introduce flexible, highly precise and fast microassembly methods. In this paper, the control system of a microrobot-based microassembly desktop station that has been developed at the University of Karlsruhe, will be presented from the lower to the planning levels. This comprises vision-based closed-loop control, user interfaces, a re-configurable computer-array, execution planning and assembly planning algorithms tailored to the needs of the microassembly station.

One of the most interesting and probably more difficult challenges in developing Software Systems is the modelling of their evolutionary capacity, that is to say, to gather the possibility that the Software Systems will go ahead in the future with the necessary changes to adapt to the environment using a different and new functionality. Modelling this evolution requires to have into account what kind of changes and modifications could follow and support a Software System during its life and also during its development. The evolutionary characteristics of a Software System can be approached by abstract evolutionary models, which can be further formalised. This formalisation makes operational the abstract evolutionary models and allows a kind of representation of the evolutionary process that could support the specification and mapping into concrete specification and implementation tools. These tools further allow us to obtain concrete and functional Software Systems.

Software systems evolve over time. Traditional software development methods and tools support partial aspects of this evolving process. Over the last few years we have researched into theoretical evolution search models which might be applied to the development of software systems. The Theory of Systems and biology have interesting views on the evolution process, very different from genetic algorithms, which may aid in the development of software systems and CASE tools. Our aim in this paper is to present a first version of a tool (HEDES) which includes these models and implements them in an object-oriented language (VisualWorks 3.0), using first-order temporal logic as support. In addition, some lessons learnt in the development of these complex and changing tools will be outlined, especially the importance of an iterative lifecycle in object-oriented development, a high degree of cohesion of the develop- ment team and the need for flexible and rapid ways of communicating new ideas using discussion meetings, Internet facilities and development support tools.

Data distribution is a crucial problem affecting the cost and efficient use of these systems. The problem is further exacerbated by the lack of methods and support tools for the design of distributed databases. This paper outlines some of the main techniques currently used for data distribution, such as vertical partitioning and replication. Two vertical fragmentation methods are described, the classic NAVATHE method and the newer FURD method, as well as the two proposed in this paper, the FURD-FDEZ and the FURD WITH REPLICATION methods.

The article is about the management of an important part of the development process of a software system: the phase of incremental implementation. Implementation is based on the design of the system architecture and its components, its main task is the production of modules and their integration.Project management has to plan and control the flow of project activities. Besides logical dependencies between them it has to take into consideration particularly the aspects of time and availability of qualified resources. In addition, each calculated plan depends upon several restricting constraints, which are influenced by real time decisions of the environmental system.Therefore a model of such a supporting management system is dynamic and non deterministic in its nature, it needs the ability of adapting itsself due to external decisions.

As we try to fulfill the requirements upon quality of the software products we cannot avoid the use of the complexity metrics. There is a lot of different metrics and also there are hundreds of tools for analyzing the software with some of those metrics. However, since all tools available are concentrated only on some specific programming metrics, for a comprehensive analysis one has to use a lot of different tools. We wanted to derive an environment that would include all of the mostly used metrics, and since we are also developing new metrics ourselves, we developed a tool called Software Complexity Analyzer, that beside classical metrics incorporates also more universal fractal metrics.

Presynaptic Inhibition (PI) basically consists of the strong suppression of a neuron’s response before the stimulus reaches the synaptic terminals mediated by a second, inhibitory, neuron. It has a long lasting effect, greatly potentiated by the action of anaesthetics, that has been observed in motorneurons and in several other places of nervous systems, mainly in sensory processing. In this paper we will focus on several different ways of modelling the effect of Presynaptic Inhibition(PI) in the visual pathway as well as the different artificial counterparts derived from such modelling, mainly in two directions: the possibility of computing invariant representations against general changes in illumination of the input image impinging the retina (which is equivalent to a low-level non linear information processing filter) and the role of PI as selector of sets of stimulae that have to be derived to higher brain areas, which, in turn, is equivalent to a “higher-level filter” of information, in the sense of “filtering” the possible semantic content of the information that is allowed to reach later stages of processing.

This paper shows the interaction among the algebraic-analytical data fields theory, the complete descriptions that may be truncated for practical visual tasks and the different level of processing in the early visual pathway, in order to establish a measure of the representation of similarity between images. Our work is based on the approach of modeling the representational capabilities of systems of receptive fields found in early mammalian vision. It is well known that a representation scheme with a metric in the space of representation induce, in a natural way, a similarity measure between images. We propose in this paper to compare different representation schemes, based in previous theorems about completnes on data fields, according to the induced similarity measure.

In this article an original formulation of McCulloch’s Artificial Program II is presented. A general methodology in order to build layered distributed granular computing machines is obtained. We apply this Program II to the classification stage of a pattern recognition system leading to Lastres Approach.

Ontologies provide an explicit and shared specification of the domain knowledge in some field. With the objective of facilitating the integration of case-based reasoning, rule-based reasoning and patient databases, we propose a medical ontology, which is inherent to the ophthalmology domain, but it can be reused by another medical domains with the same representation requirements. In order to achieve this degree of reusing, the ontology was designed making a difference between core and peripheral concepts, domain-specific and method-specific concepts and task(method)-relevant and task(method)-specific concepts.

The paper is concerned with static uncertain systems described by a function or by a relation. Unknown parameters in the mathematical models are considered as so called uncertain variables described by certainty distributions given by an expert. Two versions of the uncertain variables based on two versions of uncertain logics are defined. In the second part of the paper the formulations of the analysis and decision making problems adequate to the description of uncertainty are presented and the general procedures of the problem solving are described. Simple examples and an algorithm of the decision problem solving for a discrete case illustrate the computational aspects of the approach based on the uncertain variables and the possibility of the application to computer-aided analysis and decision making.

In this paper, a direct sliding control learning strategy is combined with a fuzzy system with variable granularity to achieve different precision requirements in different zones of the state-space; non-local basis functions are also added. A coordinate transformation gives a clearer meaning over performance evaluation and the fuzzy system operates on this transformed space.

The paper presents a set of new non-parametric and parametric identification algorithms oriented to using within the input/output system description. A purpose of elaborating the identification techniques is to involve as broad as possible, to some extend, classes of stochastic system descriptions, both linear and nonlinear ones, assuming that the lack of knowledge with respect to the system may vary from unknown system parameters to unknown system structure at all.

The paper presents examples of artificial neural networks approach for analysis of gas sensors responses. The research focused on quantitative analysis of gas mixtures appearing in dry and humid air. Despite difficulties in development of selective gas sensors, application of neural networks as self tuning signal processors provide construction of sensor systems capable of reliable measurements as well as analysis of gas mixtures with reasonable accuracy. Possibility of implementation of neural processing in low-cost devices enables eventual fabrication of microsystems integrating gas sensor matrices with intelligent data processing devices.

The scenario treated in this paper concerns the discrete event control and supervision of two decoupled continuous systems. The theoretical framework is based on the hybrid control system architecture, developed by P.J. Antsaklis and his co-workers and on the supervised control concept, proposed by the Sylodi Group from Grenoble. The main contribution of this paper is an algorithm for building the discrete event approximation of a continuous system with unknown, constant and constrained parameters. The Antsaklis formalism is extended to a disturbed continuous first order system.

One of the most important fields of research in computer science is the investigation of intelligent agents for the web which should search in the net for informations. Usually these agents are cooperating and form a working group of “multi agents”. But these groups or different agents can also stay in competition to each other. In both cases we have to model the behavior of the agents; i. e. to program a set of rules (consisting of preconditions and actions) which tells every agent what he has to do.