First European Simulation Congress ESC 83

Systems analysis provides large volumes of data. For more complex systems, the systems parameters which are important for the aim of investigation, must be obtained from this data. It is sensible to decompose the system into subsystems according to its structure, so that a well organised hierarchy has been obtained in which the individual subsystems and their relations are defined. The quality of decomposition depends on the extent to which the interactions between the submodels have been minimised. This paper describes how the subsystems are individually modelled and step-by-step combined via defined interfaces. The inter-face between each subsystem and its enviroment consists of system inputs which influence the subsystem from outside and subsystem outputs which influence the enviroment of the subsystem. The behaviour of the environment can be represented by the “environment model”, which can be replaced by tables of values or analytic functions. The submodels can be implemented by various methods, provided that the specification of the interface has been adhered to. The combination of analytic and simulation models has prooved satisfactory in practice.

The effective use of simulation as a methodology for problem solving can be ameliorated by the introduction of simulation based inquiry systems. These systems comprise a language system, a knowledge system and a problem processing system.A family of simulation based inquiry systems can be implemented in the programming language Simula. We identify the instruments of a conceptualizer, a modellor, an experimenter and an evaluator. Various entity-categories present the building blocks for simulation methodology in specific appication area’s.

Approximation-algorithms (heuristics) are important in time-critical ap- plications. The selection of a suitable heuristic depends on the expected error. Theoretical results for approximation-algorithms mainly deal with worst-case-estimations of the error. By simulation of the application-environment information about the error-behaviour of the application data can be obtained. The influence of parameters on the error-behaviour of different heuristics for the Travel ling-Salesman-problem have been analysed by this method.

A process can be described either by a physical model or by a mathematical model. The mathematical model can be based on chemical/physical laws or statistical principals. It must be clarified, whether the process can be divided into component processes or if it is nessesary to consider the process as a unit. Presented here is a mathematical model based on chemical/physical laws for the principal item of a plant for enrichment of the aqueous acetic acid and the connected waste water treatment. It will be simulated on an analog computer system. The operating costs will be determined by the objective function. Finally the plant conception will be verified.

A concept of a simulation system is developed that is provided for the assistance of the design in engineering. As a special case the design of an alternative type of a series of machines, for instance a series of centrifugal pumps, is considered. The simulation model will consist of three parts, one comprising relations for the whole series in mathematical formulations or equivalents of these, the second part giving design details of the machines that already exist, and the third specifying the alternative type more or less exactly. The procedure of using such a simulation system and the way of combining it with methods for the computer aided design is described.

The concept of sequential systems is a mathematical notion for arbitrary systems which excludes parallelism. It has been used by the author for modelling for example a system modularization concept.In this paper we describe that system concept and the system construction concept, which corresponds to the development of programs by stepwise refinement. A system on a refinement level is called a realization of a version preceding it in the construction process, the latter an abstraction or a simulation of the former. We hope that this concept will have useful applications in simulation theory.

In our lecture we summarize user demands to general simulation system, then we show, how far recently developped simulation languages satisfy these demands. We outline an idealized software system for general simulation system, and mention some theoretical and practical problems to be solved for future development of general simulation systems.

Based on the description of ecological processes by chains and cycles of rate-coupled growing systems a structure design principle is introduced which allows to describe an arbitrary nonlinear instationary system governed by ordinary differential equations as system of Volt’erra equations (Volterra-representation of the system). By a nonlinear trans-formation the Volterra-system is changed into a system of differential equations with only product terms (Riccati-representation). This unified system description in Volterra- and Riccati-representation allows to simulate an arbitrary system using only basic macros of a simulation language. Furthermore the dynamic behaviour of a system can be analysed using well known methods for analysing Volterra-equations or certain growing modes of the Riccati-representation. Two examples, the Lorenz attractor and a Volterra-Lotka integro-differential equation demonstrate this concept using the simulation languages ACSL and HYBSYS.

Complete models for mulstistage systems with a large number of stages cannot be used in process control, because of the long computation times. Even off-line simulations are expensive to run and accurate reduced order models can speed up the design of control systems. A simple, general and direct method for order reduction of multistage systems is described. The coupled ordinary differential equations, describing the process dynamics, are solved by a weighed residual method based on Discrete Orthogonal Legendre Polynomials, since the space coordinate is a discrete variable (stage number). An important asset of the method is that the state variable approach is retained and a simple relation exists between the original and the reduced model. Very satisfactory models are obtained with order reductions by a factor of 4 or more. The approximation error decreases much more rapidly with increasing order of the reduced model than with other order reducing methods.

The knowledge of a good mathematical model is an important prerequisite to the successful application of modern control techniques. Many real world processes, however, can not be described by one model with one set of parameters, because they have different dynamic and static characteristics under different operating conditions. The paper presents a simple modification of well-known identification and simulation methods so that a single algorithm can itself estimate and simulate different sets of parameters for different operating conditions.

This paper deals with the modeling and computer simulation using original axiomatic system. This system contains 6 axioms built upon two basic concepts: data and experiment. This system is presented especially from the aspect of simulation model building, becouse this systems was conceived having in mind the possibility of its use in practice. The major interest is in modeling complex phenomenous and the possibility of quantitative analysing processes of systems, and the possibility of using this axiomatic mode for solving problems arising in simulation and modeling.

The simulation of events in discrete systems for conflicting situations (i.e. race conditions, parallel processes, active feedback loops) is discussed. After a brief survey of the origins of such instances the quasideterministic state description method and the concept of the duality of discrete events are introduced as means for solving the above problems. The categorization of undetermined situations into two subclasses, i.e. Indeterminacies of First and Indeterminacies of Second Order, takes care of the state representation in “level type” events, and systems characterized by the movement of entities can be transformed into such because of their duality.

In this paper a parallel processor with a MIMD-structure, the so-called Delft Parallel Processor (DPP), will be presented. The DPP, which has been developed at Delft University of Technology, is operational since spring 1983. The architecture and the developed software for programming the DPP will be discussed. The last part of this paper will be a reflection on the general thoughts about a hierarchical hardware/software computing system with a distributed MIMD-structure.

Compact normal operators on Hilbert spaces are regarded, they are resolvable in weighted sums of projections. This implies that - once given the spectral coefficients - the “operator” simply reduces to a number of multiplications of the spectral coefficients by the corresponding eigenvalues. A parallel computer system is presented consisting of system elements of a digital differential analyzer. They are best suited because they are capable of performing those basic operations asked for when resolving operators in weihted sums of projections: the Stieltjes integration and the constant- multiplication. The interconnection structure of the system elements is the same fixed one for all the different operators of the compact normal type. The peculiarity of a definite operator is simply pinpointed by means of information, i.e. the eigenvectors and eigenvalues. This makes the parallel computer system an easily applicable tool for simulation purposes.

List processing plays a central role in simulation programs. It Is apparent that a distributed list processing is essential for a distributed simulation. The list processor reported here is a special microprogrammed processor, according to the co-processor concept, whose architecture 1) is optimized for the execution of list and arithmetic operations, and 2) support distributed list processing on several list processors. In order to accommodate a wide range of list applications a comprehensive basic instruction set has been implemented which allows the execution of many desired list operations, e.g. complex searching and sorting algorithms. For a specific application (e.g. event list manipulation) a high level instruction set can be built by using these basic instructions. For attribute manipulation an arithmetic instruction set (floating point) is also included.

This paper describes a new multiprocessor that has been developed by Electronic Associates, Inc. for scientific analysis of dynamic systems. Proven parallel and sequential computing methods are integrated in Simstar (TM) to provide a unique capability for mixed continuous/discrete system simulation and signal processing. In contrast to the earlier manually programmed analog and hybrid computers, Simstar is a completely automatic device driven from high level software in a Host data processing computer.The requirements for dynamic system simulation in different fields is developed in the paper to establish a criteria for measurement of Simstar performance relative to alternate computers. New concepts in system architecture, component technology, and system communication features are described. An overview of the Simstar programming system is covered to see the flow from the simulation language input to the program segments for the various processors. Also, a brief discussion of program operation from the Host terminals through a run-time executive is described.

Current and future military aircraft projects are making increased use of airborne digital computing, particularly in the areas of flight control, cockpit displays and weapon aiming. The Development Simulator at Warton, which includes an air-to-air Combat Dome and an air-to-ground cockpit, has recently updated its computing facilities to cater for the future requirements of pilot-in-the-loop simulation. The new system architecture is centred on four Array Processors which provide the high speed computing required to achieve representative aircraft and undercarriage dynamics. Two multi-port shared memory systems provide a communication network for up to 12 separate PDP 11 processors with minimal system software overheads.

The AEEW analogue computer Installation contains three EAI 680 computers. These are used, to study the dynamics of nuclear power plants, to test the performance of control hardware and for the training of reactor operators- Problems studied on these machines are prepared with the assistance of the APSE (Automatic Programming and Scaling of Equations) digital computer code. (Ref 1). This paper describes the application of microprocessors within the analogue models, which has significantly reduced the analogue equipment requirement and Improved both accuracy and capability.

A simulator for a multiprocessor system consisting of p identical processors and m shared memories connected by a crossbar switch is presented. Its main features are a DO-loop unwrapper, the scheduler and the emulator. Special attention is given to the automatic partitioning of a high level program into parallel tasks. This allows the user to concentrate on different processor configurations and experience with many algorithms. The simulator takes into account communication delays and variable execution times for different operations.

The hybrid time-sharing system MACHYS is presented. It has been developed at the Hybrid Comp Centre of the Technical University in Vienna and is fully operational since September 1982. Up to 8 hybrid users can have simultaneous access to all hybrid computing resources, either from terminals or remote telephone links. The mathematical building blocks of a fully equipped EAI 680 analog computer are controlled by high level software commands in an machine independent way, using an electronic analog switch matrix and a self developed parallel logic processor. The interactive compiler HYBSYS supports a problemoriented, block structured model definition. Simple commands allow optimal online scaling, grafic and alphanumeric documentation as well as sophisticated experiments with the model like parameter optimization and nonlinaer zero finding.An improved time-sharing algorithm, the results of some measurements and typical problem change times are presented.

This test bench was designed and implemented in order to evaluate newly designed sampling systems or systems which are already used in practice. The simulation package is a special purpose system in order to guarantee a user friendly man-maschine interface, espescially with regard to users in industry with only medium knowledge of statistics and simulation. The package consists of the four modules: CONTROL,HISTORY,INSPECTION and STATISTICAL-ANALYSIS Br REPORT-GENERATION.The system may be run either in a batch or in an interactive mode. It modules are written in FORTRAN 77 and are well structured prgrammed. It is not fully portable due to the zero layer level which is written in a JCL and the the I/O functions used.

Based on and extended toConputerDesignLanguage andInstructionSet.ProcessorSpecification, the language CDIM is proposed as a system both for design verification at different levels of abstraction and for dynamic verification of design refinement steps. Some aspects of structure, behavior and algorithm description level and design decomposition are discussed.

Ship’s design and more especially the simulation and modelling of all the relevant aspects connected to the real-life functioning of a ship is increasingly becoming a multi-disciplinary affair. Naval architects, port-experts, economist and mathematicians are all busy with the mono-disciplinary optimization of design and performance. A meaningful link however between these activities seems to be very difficult to be realised: all the experts talk their own language and are unable to understand and integrate the often very relevent information from the other disciplines in their own work.An interdisciplinary approach to the problem as a whole could, in some cases, result in a distinctly positve synergy.Appropriately designed simulation methods and techniques are needed in order to guarantee the success of this interdisciplinary synthesis.

SDS (Simulation of Dynamics of Systems) is an interactive simulation language intended for the Discrete Simulation of models that can be shown by systems of algebra formulas. These models are apstract, dynamic with a discrete modification of state and with uniformed-discrete change of time, with deterministic and stohastic connections that can be linear or nonlinear. The simulation language SDS is especially suitable for the simulation of dynamics of systems that can be described by elements of state, elements of modification of state, and flows of matter, enery and informations. The main characteristics of the language are: it is easily learned and simple for use, it has a high degree of cimilarity between the program and the simulational model, the programming is simple and it is not conditioned on the knowledge of programming technics, it assures scemantic and synthetic control, already in tite procedure of creating the program, it automatically defines the position of instruction in cue of execution, it assures, by use of external data set, the independence of program from data, it makes interactive work of more users at onle possible, it is of modular structure, so that the use of external program packages, there is a possibility of widening on behalf of the user, it demands modest computer resources, it is independent of the operating system of computers, it can be installed on most of the existing computers.

EARLY DESIRE (DirectExecutingSImulation inREal time) is a new floatingpoint equation-language Tor interactive continuous-system simulation on DEC PDP-11 or LSI-11 workstations. Programs run 1.3 to 5 times faster than threaded FORTRAN and execute directly on a typed RUN command; there is no visible compiler or linker. An interpreted job-control language provides interactive program entry, editing, and file operations and programs multi-run simulation studies. The “dynamic” program segment determining the simulation model is entered just like the job-control program and accesses the same named variables A small, very fast compiler translates the “dynamic” segment practically instantaneously before the first simulation run. Different precompiled integration rules may be overlaid from mass storage.EARLY DESIRE can handle up to 40 state variables and provides a true CRT screen editor with cursor-control keys, a macro facility, and run-time graphics even on inexpensive alphanumeric terminals. The language admits subscripted variables and permits especially efficient simulation of sampled-data control systems. Future DESIRE systems will be written in standard PASCAL for machine independence and will run on 32-bit superminicomputers.

The program package IMP (Interactive Modeling Program) is a tool for interactive model development and simulation. It is illustrated how to work with IMP. Moreover the input of variable parameter values via potentiometers, the graphic output and the internal structure of the program are explained.

Simulation is getting more and more Important for designing and developing systems. In the software design simulation is not very much emphasized despite of the complexity of the problems to be solved. In BORIS, a block-oriented interactive simulation system for discrete systems, the simulation modeling process conforms with the software design methodology. Therefore BORIS is most suited for modeling and simulating the software system and is applicable to the software life cycle. The simulation model corresponds to the structure of the designed system; it is realized as a net of agencies. An agency is a functional block with inputs and/or outputs. The agengies are linked to a net of agencies. Keywords: Software Design Methodology, Software Tool, Modeling, Simulation, Agency, Net of Agencies, BORIS

One way to calculate the behavior of computer networks is simulation. It allows to model nearly all problems but it may be rather difficult to prove the correctness of the solution. Therefore the results must be validated by measured values for example. On the other hand an analytical solution that is easier to validate exists for networks that base on Markov processes. A subset of them is characterized by a ‘product form solution’ that is supported by a lot of computational algorithms and that is applicable to a large number of problems. The paper deals with open networks that involve summations of infinite numbers of items. It presents a modified algorithm for an exact or an arbitrary well approximated computation of important performance measures. In order to decrease the computation time without loss of accuracy, special properties of the probability density functions, state aggregation and network decomposition are used.

COSMOS is a new simulation language for description of continuous, discrete, or combined continuous discrete models. COSMOS coded continuous simulation programs may look quite similar to CSSL type language coded programs. As a matter of fact, the COSMOS language definition even constitutes a superset of CSSL since many additional structuring elements are incorporated, e.g. for comfortable modelling of piecewise continuous models and variable structure systems. COSMOS coded discrete simulation programs look similar to SIMULA programs, and not similar to GPSS programs. In addition to the continuous and discrete simulation features, newelements have been incorporated into the language definition to allow for coding combined continuous and discrete systems as well. The model and experiment description are totally seperated. A powerful set of experiment description statements is available.

The aim of this paper is to compare the features of the dig-tal simulation language ACSL and of the hybrid simulation language HYBSYS (developed at the Technical University Vienna) for the simulation of regulator plants with alternatively continuous and discrete feeeback. Proposals for the implementation of a discretised feedback are given. As example the simulation of the disturbance rejection control of a magnetically levitated vehicle is considered.

FORCASD (FORtran-based and evaluation net-oriented Computer Aided System Design tool) is a portable program package that supports modelling and simulation by evaluation nets. As a special feature, a low-cost graphic display with a light pen facility can be used for model definition by drawing evaluation nets interactively on the screen. The graphical images are automatically converted into appropriate FORTRAN program parts representing those evaluation net structures and form the essential parts of the simulation program.

In the present paper we are discussing the language Ada., We are giving a short survey of the history of Ada and are presenting the main design goals of the language. We compare Ada with some other general programming languages and with specific simulation languages. We investigate the advances of Ada for a simulation support environment in form of a package system. We analyse Ada in the context of different simulation approaches or world views. Finally we are discussing, how a simulation language could be defined on top of Ada and are giving a simple example for such an extension.

The SYSMOD language represents a complete departure from the more traditional continuous systems simulation software,(in the main ‘67 CSSL based /STRA67/). The change was motivated both by recognition of the trends in simulation and by a more general observation of the developments in the conventional programming language forms and facilities. SYSMOD is a portable language designed predominantly to support the continuous modelling formalism as applies in the aerospace application field. Necessarily, however, it has an extended capability in support of both the discrete time and the discrete event formalisms. Such extensions enable its categorisation as being of the “combined” simulation class. Within this paper the structure of the language is introduced in the context of the underlying rationale of the facility inclusions.

The mechanisms for flow control for queued systems in the simulation package SPIRO are described. These can be regarded as consisting of four discrete levels: control of server networks, control of servers, control of stores, and interaction between tokens. Examples are given for each level.

In models of queued systems queues of objects form at stations. We describe, with examples, the stations supported by the simulation software GPSS, GPSS-F, SLAM, and SPIRO.

A68SIM is a program system based on ALGOL 68, containing a set of procedures and operator definitions, which are designed to aid the writing of ALGOL-68-programs for discrete event simulation. A language with the power and flexibility of ALGOL 68 provides a variety of means for carrying out simulations, and A68SIM is just one of the possible approaches. A68SIM is a set of basic elements, and proceeding from these elements the user has a lot of opportunities in formulating and solving his simulation problem. This paper describes some features of A68SIM, originally developed by D.C.S. Shearn, Division of Economic Studies, University of Sheffield, U.K. and since some years implemented on our CD CYBER 172 and CYBER 76 at the Regional Computer Center at the University of Hannover, and compares it with SIMULA 67 and GASP IV, two other simulation “languages” at our site.

This paper reports experiences with the conversion of a radar system model written in FORTRAN to a formulation in a continuous system simlation language (ACSL). The mathematical model includes geometric equations, noise sources, transfer function blocks and a logic to control model execution. All conversion steps are described and practical aspects like efficient computer usage or readability of model code are discussed. The report ends with a brief outlook to future modeling development.

With conventional simulations (e.g. using the well-known simulation languages SIMSCRIPT or DYNAMO) coefficients of difference equations must be calculated by means of common statistical procedures implying a reduction of the amount of information. A certain loss of information is generally accepted. The MAPLIS language provides one- or multidimensional matrices (tabulations) instead of scalar coefficients. No reduction of information is required since the results of cross tabulations can be used immediately. Furthermore, a model designer may decide how many dimensions he needs. Some examples taken from the system of models as being used in the so-called “Münchener Simulationsmodell” (microdemographic prognosis) show the characteristics of the MAPLIS language: in form: how the various models and variables can be declared; how relationships between variables and between models are established. In content: how correct designing and programming of models is supported by a compatibility check of variables acting to each other; how dialogue with the computer enables the user to maintain, run and control models, to simulate exogeneous events by means of modification, which is not restricted to data: existing models can also be replaced during the run and new ones may be added.

Some 3-body-systems of seemingly similar movement (balls of a juggler, a bus service on a scheduled round route, and 3 racing cars on a track) are described in terms of states and events and modelled with Petri nets and P-Ta-nets.Stress is laid on the difference between the static system components (including their structural relations), and the dynamic processes (being deterministic or decision-dependent) which run on the systems. Both of them have to be mapped adequately to the simulating model. A Gedanken-Experiment shows that the difference in control is the reason for the difference in results and in the reliability of results which are obtainable from a simulation.This leads from the simulation problem to a modelling problem, which overrides the actual (numerical) simulation in importance. The special problems of modelling distributed and competitive systems are shown and some solutions (nets) offered.

Resource sharing in multiprocessor systems has a strong influence on the speed-up of the processing elements. Analytic computational methods, e.g. queueing theory, do not provide time-dependent solutions by tolerable effort,’ whereas simulation often requires a great deal of computational time and is more difficult to validate. In order to overcome these disadvantages, the paper presents a model that combines analytic computation and simulation. The time-dependent solution of the model under time-dependent load is obtained by the simulation package FORCASD. For reasons of state reduction and decrease of computational effort the hardware layer is modeled by a queueing model. The results are validated by measurements in the M5PS multiprocessor system.

Experiments based upon computer simulations of engineering control systems can provide students with an introduction to interesting problems of analysis and design and can complement more conventional analytical and computational work. These simulation-based studies also complement the traditional type of laboratory experiments found in most control engineering courses and extend the range of control systems applications to which students are exposed. This paper describes an analogue computer simulation of an aircraft lateral beam guidance system and gives an account of its use within a control systems course for undergraduate students of electronics and electrical engineering.

Quick transportation of containers by truck and train is one of the expanding markets of good transportation. The basic idea of a new method of container transfer is the following: Containers shall change from one goods train to another as people change from one passenger train to another. This procedure is done by special rail-bound vehicles, which transfer containers from one train to another. The container transfer is supported by an operating system BFS which consists of the operating computer BFR and the hierarchically organised software of the BFS. A simulation system which serves as a substitute for the container transfer station allows the realtime simulation of the complete container transfer process. Petri-Nets have been employed to derive the description of major parts of the simulation from the parallel structure of the process. A second revision of the simulation system and the operating system has been completed.

The computation of railway dynamics on an irregular track leads to strongly nonlinear equations, which are solved in the time domain by means of a hybrid computer Yet the modelling of a single wheel is very complex, due to the nonlinear shape of the profiles and due to nonlinear saturation effects of the creep forces in the contact area. Without developping new simulation techniques, the analog simulation of a whole vehicle (with eight wheels) would require an excessive amount of computing-components. This problem is successfully solved by modelling the characteristics of only one wheel on the analog computer, and using it for the dynamic simulation like a subroutine call in a digital program. Two bogies with a total of four independent wheelsets are simulated by using a multiplexing technique, based on the logical circuitry of the analog computer.A digital computer is interfaced to the analog machine in order to vary constructive parameters, to output the track irregularities, and to digitize forces and essential state variables during the process of simulation. These data are used for the evaluation of quality criteria (such as ride quality, lateral force level, wear characteristics) and as input for an optimization algorithm.

Controlling a technical process only or mainly using solar energy makes specific problems. One main problem is that solar energy varies over the time. So, partial processes or components are to be activated in a way which is optimal not only with respect to their sequences, interconnections, and parallelisms but, furthermore, with respect to their energy requirements and the actual energy supply. This paper in particular deals with the example of a brewing process using solar energy which is projected by VDP1 and promoted by BMFT. It will be sketched how some of the problems can be solved using net models and dynamic simulation techniques.

A simulation method for investigations of transient behaviour of queueing systems is described in this paper. The method is based on a modification of the event-by-event simulation technique where arrivals according to time-dependent, generalized Poisson processes are considered. Applications of the method to performance evaluation of switching systems are discussed where the transient behaviour of the M[X] (t)/M/1 queueing model is investigated. The more general queueing system M(t)/G/n and its time-dependent responses to instationary traffic patterns are also considered in the last part of the paper.

A mathematical model for lead acid batteries is established in bond graph form. The causal assignment procedure is used to determine input-output interactions between components. Direct programming of a continuous system simulation language from the bond graph is illustrated.

A case study of disturbances on a high frequency measurement set-up is made. The sources of error are carefully analysed and modelled. A difference in treatment is made between stochastic and frequency dependent systematic errors of the network analyzer, also reflections and missmatches are taken into account. Validation of the whole error model is possible by checking a physical constant or a known parameter.

This paper applies to a certain type of crane used on oil rigs in the North Sea. A 7th order system of differential equations is established to represent the hydraulic system and the motion of the crane structure and load. First order approximations allow the structure/load combination to be considered as a linear double mass-spring system where the structure stiffness and mass depend on angular position of the boom. The validity of these approximations is confirmed using computer simulation. Further assumptions allow decoupling of the linearised 7th Order system so that the frequencies of the modes of oscillation can be calculated from system parameters. The danger of resonant instability can then be assessed.

By using partial elemental balances a chemical balance representation of the growth process could be implemented on a process computer for monitoring biomass. Due to the fact that the elementary compositions of biomass, substrate and in some cases of products, are not always known, one has to introduce inductive information in the proposed model. A sensitivity analysis will show the impact of this information on the ultimate “net-effect” sensor. Finally, it is hoped that the paper will bring some insights in the use of model activities in sensor technology for computer- coupled fermentation.

With the help of the digital simulation the trial has been made to simulate a narrow and broadband multisubscriber dialogue system, hereby a real system, the VBD-labor system, served as pattern for the model structure and its verification. The simulation model, developed in the Heinrich-Hertz-Institut, Berlin, is comparatively quickly adaptable to other configurations by the parameter control. Thus, e.g. by parameter input the number of computers, the processes in the computers or the number of external storages may be modified. Inspite of the complexity of the simulated system and the high specification in the model (task level) the computing times in case of simulation runs on a timesharing equipment have been not higher than in case of real runs. By the trials simultaneously made within the project at the realized multisubscriber broadband dialogue system the function of the model could be tested. Consequently, it is assured that the complexity of the model is sufficient to introduce it also when planning new systems.

Implementation of multivariable control using microprocessors or signal processors introduces several effects due to nonideal behaviour of the digital controller. Such effects like quantization or sampling delays can often not be neglected, particularly in applications with fast control systems. A discussion is given of possible nonideal behaviour of digital controllers. Software modules to simulate digital control are embedded in a simulation concept which combines some block oriented description with centralized integration.

Es wird ein Modell für die Magnetisierung von Eisen hergeleitet, wobei die Bewegung der Blochwände auch berücksichtigt wird. Dies führt zu einer par- tiellen Differentialgleichung (des „Wärmeleit“-Typs) mit sich bewegenden Unstetig- keiten (Blochwänden). Dieses Problem wird als „Zwei-Punkt-Randwertaufgabe“ mit Zu- standsereignissen betrachtet.Die durch Simulation gewonnenen Resultate bestatigen die Notwendigkeit, die Bloch- wandbewegungen bei dynamischen Magnetisierungsvorgangen im Eisen in Betracht zu ziehen.

A modular, all digital, six degree-of-freedom missile simulation provides a flexible tool for the investigation and analysis of alternative missile system elements such as optimal guidance laws, autopilot design, seeker/sensor systems, propulsion systems and airframe configurations. Each module of the program describes either hardware components (e.g., seeker, actuator, etc.), homogenous environmental conditions (e.g., winds and atmospheric data), or simulation-unique functions (e.g., kinematics, terminal geometry, etc.). This modular approach facilitates adaptation of the program to reflect changing hardware components and different levels of environmental modeling.To illustrate the utilization of the replaceable missile subsystem modules, two different guidance laws for an air defense missile are analyzed using the simulation. The first guidance law, proportional navigation guidance, was used as the baseline guidance law in the simulation because it was simple to implement. Its main drawback is its sensitivity to high end game maneuvers and external disturbances. The second guidance approach is noisy discrete-time disturbance-utilizing control which is formulated as a form of modified proportional navigation. It delivers improved miss distance performance in scenarios which include high levels of maneuvers, winds and other external disturbances. However, the implementation of the disturbance-utilizing control guidance is more complex because a real-time estimation of kinematic states and time-to-go to intercept is required.The application described in this paper demonstrates how the modular simulation approach provides an efficient analysis tool for conducting trade-off studies between two guidance techniques requiring essentially different implementations of measurements and estimates. Particular emphasis is placed on isolating the scenario conditions under which the more complex implementation is needed.

The DC link converter is a viable option in the design of continuously variable transmission in, for instance, hybrid vehicles. To investigate the feasibility of using natural commutation, a model is required in which strategies for link current control may be tested. A simulation method is described which involves the integration of the equations of state of a synchronous machine with a discontinuous unbalanced load.

A frequency domain simulation technique has been developed and applied to the investigation of the performance and of the losses of inverter-fed electrical drives. Components of the drive system are calculated as two- or four-terminal networks and described by their frequency response function. In order to analyse the stationary behaviour, all components have to be calculated for each harmonic considered. The results are superposed and presented in time domain by use of Fast Fourier Transform. This simulation technique and the simulation results, compared to measurements of a test drive, are presented here.

Electrohydraulic servo-drives are used favourably in manufacturing industry because of their great actuating power. Their distinct nonlinear behaviour as well as their characteristic weak damping, however, complicate the control of such drives considerably. Subject of this contribution is part of an investigation of the design of linear feedback controls for servovalve-controlled hydraulic drives, that is based on a hybrid simulation technique. State feedback controls with observers as well as multiloop controls are tested and results of the simulations are presented in this paper.

In contrast to many other industrial processes the design of control systems for air-conditioning plants cannot be based on fixed operating point parameters. For the design of control systems it is necessary to investigate the functional interrelations between operating point parameters and the dynamic behaviour. Due to the high costs of measuring the dynamic behaviour the investigations to be described are carried out by digital simulation. Based on simple mathematical models, which are available in the literature, the continuous-time transfer functions of an interesting configuration are transformed to a linear discrete-time description using the Z-Transform. For this step the structure of each device is taken into account. The results are a basis for discussion of the necessarity and the practicability of modern controlled systems in air-conditioning plants.

Two mathematical models for the study of the thermal state-in the exhaust well of an underground gasification plant are set up: a distributed parameter model and a reduced-order model of fifth order. Computer requirements and simulation results are presented for both models and compared. It is shown that the reduced-order model retains the major nonlinear and dynamic aspects of the complete model, and will be quite satisfactory for investigations on the control of the plant.

Animals with total artificial heart replacement (TAH) have survived for up to 9 months with manually controlled driving systems. A certain physiological control is provided by the animals intact peripheral regulation resulting in pathological venous hemodynamics, expressing the needs for a technical automatic control system, replacing the natural sympathetic stimulation and Starlings Law of the heart. A nonlinear dynamic mathematical model of the human cardiovascular system (CVSM) containing the systemic and pulmonary components was outlined and verified, which is employable to analyse the hemodynamic behaviour under normal and abnormal states. By modifying this model, assuming that cardiac output is constant at a preset level, the hemodynamic behaviour of a circulation system with an artificial heart was simulated. Comparing the simulation results with those from TAH experiments, a good accordance in the response of right atrial and aortic pressure under exercise conditions could be proven. This gives hope, that the modified and tested cardiovascular system model can be used as a tool for computing optimal control parameters under given conditions.

The knowledge of closing ability, pressure- and energy loss can facilitate the selection of the type of heart valve to be implanted. The mock circulation has to fulfil the following terms of analogy in order to test a valve prosthesis: Geometrical analogy of the perfused parts, analogous viscosity of the fluid to that of the blood and analogy of the insteady flow proceedings. Mock circulation may be demonstrated by the examples of the St. Jude-Medical-valve. Reflux volume and pressure loss have been registred in mitral- and aorta position, and insufficiency and energy loss have been calculated. The reflux is constant and the insufficiency decreases at increasing the stroke volume. Pressure- and energy losses increase ex-ponentielly with cardiac output. At present the mock circulation is being used to test prototypes, extirpated valve prostheses, and postmortal taken heart valves.

A model is developed which describes the pathological blood pressure behaviour after treadmill test in patients with coarctation of the aorta. The model is simulated and identified using the hybrid simulation language HYBSYS. Five parameters characterize the model and consequently the effect of the coarctation after (and during) treadmill test. The identification based on 66 patients yields to interesting relations between the five parameters: as result, two parameters are sufficient to characterize approximatively the pathological blood pressure behaviour.

The possibility of developing a managable closed nonlinear mathematical model of the renovascular system for longterm blood pressure regulation is outlined. The developed model is a parametric one. Its structure contains explicitely the biological parameters.The model is studied by simulation. In comparison with experimental results, obtained from animals, it can be shown, that the developed model matches the real biological system well. Hence it is suitable for prediction and parameter estimation. This means biologically, that it is possible, to compute with the outlined model parameters, which are not measurable neither noninvasively nor in-vivo. In addition, the hemodynamic behaviour is not only assessable in normal states, also in the pathological state of hypertension, for instance essential hypertension or renovascular hypertension. Also hormonal dependent mechanisms, for instance the renin-angio- tensin-aldosteron system can be implemented.

Transient oxygen and dioxide exchange processes within the upper skin can be simulated with the model of a microcirculatory unit. After a short description of the capillary loop model the corresponding mathematical nonlinear parabolic initial boundary value problem and the numerical procedure to calculate an approximation of the solution is presented. Some results of the calculation are described and discussed representing different oxygen supply situations of the upper skin.

The influence of the arterial blood pressure (ABP) on the intracranial pulse pressure relationship (PPR) was studied on the basis of a mathematical model. Simulation revealed an opposite effect of mean ABP and ABP amplitude on the slope of the PPR. Qualitatively, the same pattern of reaction as predicted by model simulation was found by linear regression analysis in all patients in a clinical study.

For the condition of time optimality the fine control of movements like horizontal eye or head rotations is governed by characteristically patterned neural control signals.These signals mirror indirectly the dynamics of a particular motor subsystem that may be normal or changed: experimentally,through a disease,or in the model.

Profound pharmacological and pharmacokinetical studies of drugs are unavoidable in the modern realization of successfull therapy. Linear and nonlinear pharmacokinetical modelling and simulation enable dosage forms and dosage regimen design necessary for the optimal manufacturing, consumption and application of drugs. So the needful “in vivo” studies which are expensive, time consuming and sometimes delicate, can be minimized. The work deals with the analog-hybrid simulation of the concrete drug pharmacokinetics namely Cimetidine. In the six compartmental model also the first-pass effect with liver cumulation of drug, enterohepatic cycle with time changable parameter simulating food administration and corresponding time delay after the oral application of drug simulating dosage form desintegration were taken into account. The study was realized using analog-hybrid computer EAI-580 and the corresponding “in vivo” data.

In contrast with certain other areas of neurophysiology, modelling and simulation techniques appear to have failed to make significant contributions to research on the muscle spindle. In this paper an attempt is made to identify possible reasons for the failure of these methods. Suggestions are also made, with illustrative examples, of ways in which experimentation and model-related activities could be more closely integrated to the advantage of all involved in muscle spindle research.

Cross-bridge behaviour within the sarcomeres of a muscle fibre is responsible for a great deal of muscular behaviour. Because cross-bridge properties cannot be measured directly, postulation of a basic structural concept for cross bridges is needed. To confirm this postulations, a mathematical model basing on the structural concept was derived. The results of; the model were similar to those of some experiments. For the explanation of further experiments, additions to the structural concept had to be made and the mathematical model had to be reformulated until all desired experiments could be explained by the model. Using this procedure an improved model was derived, which shows reasonable steady-states, transients, chemical energy turnover, stretch response and latency relaxations.

Advantageous usage of the parallel computation properties of the classical hybrid computer in the simulation of action potential propagation along myocardial tissue forms the topic of this paper. Considerable reduction of computation time, ease of intervention into the process and live display of time courses facilitates theoretical model research in heart fibrillation.

In this paper we describe the implementation of myocardial fibre action potential generation units /BEEL77/ in the Delft Parallel Processor and the interactive experimentation on these models. This work has a double purpose: experiencing the Delft Parallel Processor on the basis of a rather complicated simulation study and, secondly, to provide colleague researchers /PASV83/ with simulation results, which should facilitate the verification of their simulations.

An n-dimensional Evolution-Operation Program has been developped that uses non-linear approximation for the evaluation and optimization of process design variables. Ridge analysis has been incorporated to obtain an spherical search pattern, which lends itself to experimental designs that are restricted to spherical bounded regions. This requires that the practicioner has in effect specified the range of exploration and approximation in which he is interested, both conditions being always met in process design. This technique has been applied to design the optimal operating conditions of a furfural production Plant reducing the experiments to a minimum, ensuring a minimal cost design. The method also allows for critical design variables decision and failure prevention.

The paper deals with the problems associated with NaOH dissolution simulation, solution of them and the applicability of the simulation scheme. First the mathematical model of the exotherm reaction in the batch reactor including some nonlinearities and mathematical model of the cooling system are given. The problems we met during the preparation of the problem for simulation were mainly the unacquaintance of one constant in the Arrhenius equation and the wide range of time constants of the reactor and cooling system. Using this approach the given problem is simulated first on the analog computer, where the chosen model can be verified and the unknown constants can be estimated by means of some experimental results, respectively. Once the model is well known, the digital simulation, using standard simulation subroutines enables the simple documentation of results and simple storage of programs. By this approach the advantages of analog and digital simulation are employed and in addition the simple verification of the used programming techniques is enabled.

An important problem for safe process control is the attenuation of temperature peaks due to exothermic chemical reactions. Steady-state simulations of a temperature peak attenuation by catalyst dilution at the entrance of a polytrop tubular packed-bed have been carried out using the reaction model of the exothermic ammonia synthesis. The requirement of the temperature peak attenuation together with a maximum yield can be met by an optimal dilution profile.

The direct approximation of a distributed parameter system by a sufficiently accurate lumped parameter system often leads to models of very high order. This dis-advantage can be overcome by an appropriate reduced model based on physical experiences. By using the example of the catalytic fixed-bed reactor the problem of model reduction of chemically reacting systems is treated. The control design is based on the concept of state control and is carried out by using frequency domain techniques. The method proposed ensures acceptable transient response characteristics of the closed-loop system as well as extensive rejection of external disturbances. Furthermore, for step-wise servo inputs, the controlled system is dynamically decoupled.

Differential equations which describe biological systems can generally only be solved, if at all, by assuming idealizing initial and boundary conditions. The treatment of such equations by simulation methods is thus of growing imprtance. Block-oriented simulation systems, which invole much less work than the writing of special programs in a higher programming language, are convenient tools for modelling continuous systems, especially with respect to non-linear components and delays. In this contribution the advantages of a particular block-oriented simulation system (improved version of MISS as described in GOTT76) using a robust stiff integrator are discussed. Here it is applied to two biological processes: one simple example from population dynamics and one from molecular biology.

The aim of scientific research in biology and medicine is to describe and - perhaps - understand structural functional relations between elements of given systems. In this survey, it is shown how to use graph theory for modelling and analyzing biological systems. Some examples of its application are presented. Every binary structural relation can be described by a graph. Calculus of weighted directed graphs can be used to model functional relations directly. Consequently, the theoretical model of a graph can give both a first insight into the structure as well as a description of the functional relations of the elements of biological systems.

The sensory feedback system of decerebrate, paralyzed cats was studied at different neuronal levels during passive movements of the animals’ forepaw. Information about the static parameters of the movement was primarily found at levels of low integration while responses to dynamic parameters were obtained at all recorded levels. However, at levels with a high degree of convergence information about momentary position was only observed at times, when dynamic parameters were also transmitted. A model is derived, which is characterized by directly processing the dynamic parameters whereas static parameters are linked with different components of the velocity signal. Furthermore the model is expanded to a n-dimensional system with an interactive exertion of the afferent channels. The results of a computer- -simulation of this model are compared with the experimental results.

One of the major factors affecting the performance of material requirements planning (MRP) systems is the uncertainty in the availability of the components required to produce assemblies and finished products. Due to the complexities of the product structures and interactions between a large number of system variables, it is not possible to use analytical techniques to assess the impact on the overall performance of MRP systems. Computer simulation represents a possible alternative. This paper describes a material requirements planning simulator (MPPSIM) designed for use in conjunction with computer-aided production control systems. Examples of the use of the simulator to assess the impact of uncertainty, when considering the use of lot for lot and fixed period requirements lot sizing techniques, are given. The effect of maintaining various levels of safety stock on the overall performance of the MRP system, in terms of the percentage of jobs started on time, stock value, work-in-progress, and cumulative production, is assessed.

This article describes a Decision Support System for modelling arid evaluating specific bankhall organizations. The DSS has been developed by the Rabobank Nederland for its organization consultants. The article pays attention to the structure of the DSS, the structure of the Simulation Model, the data organization and the Evaluation Reports. At the end some technical figures are given.

The evaluation of operating costs and supply reliability represents an important tool for the long-term system planning. Former analytical methods eva¬luate the expectations of costs and reliability indices only. Expectations however are not always sufficient for the power-system planning. Calculating standard deviations of costs and reliability indices by means of analytical techniques consumes too much time and memory in practical applications. This paper presents two methods of calculation based on Monte Carlo simulation, which evaluate the probability distributions of costs and reliability indices. Both methods differ in modelling of power-systems and consumption of computer time. Therefore the presented methods are suitable for different applications within the field of system planning.

Following the so called time-geographic point-of-view a simulation model is presented for simulating household activity projects in various locational systems. The model provides an answer to the question whether a given activity project is physically compatible with the constraints in the project itself and those imposed by the household environment consisting of household resources, like vehicles, persons etc, the facility locations with their opening hours, where certain activities can be carried out and the time-space structure and design of the transport system.

By the development of FIBS the GMD pursued the goal to provide easy to use software tools for supporting the practitioner oriented work with socioeconomic simulation models and their data. NBS shall improve the productivity and effectiveness of professionals in the preparation of their decision functions. In this paper at first the requirements of advanced decision support are discussed in aspects of data-, method- and model base- systems, decision support system generators and dialogue techniques. Then the realisation of some of these aspects in M3S is shown by examples describing how the user with little or no experience in data processing is enabled to deal with planning tasks interactively at his working place stretching from ad hoc data analyses to the use of simulation models. Finally in a scenario describing the next generation of interactive planning systems topics like group support, decision rooms, expert and knowledge- based systems are treated.

In this paper we analyze the non linear behaviour of elastic cables acted upon by dead load and tensile forces. Such cables are often employed in long-span cable-stayed bridges. We also study the stay‘s stiffening obtained by a counterstays system and used to reduce the overall defor-mability of such bridges. The analysis is made by using two non linear finite element models here proposed and suitable to grasp the geometric nonlinearity of these structures.