MLDesigner

from Wikipedia, the free encyclopedia
MLDesigner

MLDesigner.logo.small.png
Basic data

developer MLDesign Technologies, Inc, Palo Alto, CA, USA
Current  version 3.1 (December 2, 2015)
operating system GNU / Linux , Microsoft Windows
category Simulation , modeling
License proprietary
German speaking No
www.mldesigner.com

MLDesigner is a modeling and simulation tool for the development of complex embedded and networked systems . It allows the modeling of architecture and function of mixed event-discrete, time-discrete and time-continuous systems and thus the investigation of architecture, function and performance of complex systems. It is based on the non-commercial Ptolemy project of the University of California Berkeley (UC Berkeley). MLDesigner is developed by MLDesign Technologies Inc., Palo Alto, CA, USA in cooperation with Mission Level Design GmbH, Ilmenau, Germany.

Conception

Models are created in the MLDesigner in the form of graphical, hierarchically structured block diagrams. Models (systems) that can be simulated consist of modules and primitives that are connected via inputs or outputs (ports) and / or arguments (parameters, resources, memories and events). While modules in turn can consist of modules of lower hierarchical levels and primitives, primitives represent the atomic units of models. Their functionality is described by C ++ source code or with the help of finite state machines (FSM). MLDesigner has more than 2000 ready-made library elements so that programming of primitives is not absolutely necessary for modeling a system. The model elements are stored in MLDesigner in an XML dialect.

functionality

Simulations are carried out in MLDesigner on the basis of multiple calculation models adapted to the problem under consideration. A calculation model that describes how data is to be exchanged between the model elements and in which order they are to be carried out is called a domain in the MLDesigner. The various domains differ with regard to the underlying time base - a distinction is made between time-discrete, time-continuous and event-discrete model types. A mixture of model elements from different domains in one model is possible, which is why the modeling and simulation of time-discrete, time-continuous and event-discrete systems is possible with MLDesigner.

Domains

The multi-domain simulator MLDesigner supports the following domains, among others:

SDF domain (synchronous data flow)

The SDF domain is purely data flow-oriented and is used to model time-discrete systems such as B. digital signal processing systems are used. In this domain, model elements synchronously (simultaneously) exchange a fixed number of data elements, so-called particles. The SDF domain has no concept of time. The order of execution of connected model elements is statically determined on the basis of fixed rates of generated and consumed data elements (particles). Another data flow-oriented domain is the DDF domain (Dynamic Data Flow), in which, in contrast to the SDF domain, the rate of generated and consumed data elements is variable and therefore dynamic changes in the processing of data and the execution sequence can be taken into account.

DE domain (Discrete Event)

The DE domain allows the modeling of event-oriented systems such as communication and / or data networks. A model depicts the occurrence of events at certain times. The data exchange corresponds to the triggering of an event of a model part that affects another model part. The model elements are executed chronologically according to the order in which the events occur. All model elements are subject to the same global time. The modeling is supported intuitively by mechanisms such as special events or resources, which simulate the use of quantities or calculation times on processors, and are significantly simplified by appropriate pre-made model elements such as scheduling mechanisms. With the help of the DE domain, a large number of different systems, such as bus systems , digital controls , computer architectures or system on a chip (SoCs), can be modeled and examined. This includes all types of event-oriented systems. Even processes can be reproduced, simulated and optimized with MLDesigner using the paradigm of event-process chains (EPC). At the same time, the DE domain forms the core of the modeling options for architectural models, which, in connection with the modeled function, allow the investigation of the performance of systems.

FSM domain (Finite State Machine)

FSM is a concept for modeling discrete-event systems by means of state machines, whereby an FSM depicts the current state of the system or a system part, in that the occurrence of events causes a specific, modeled action and a change of state. It thus represents an alternative form of description of discrete event systems and is mapped internally to a DE model for the simulation. The modeling of FSM models takes place in the form of StateCharts . With the help of the FSM domain, systems can be modeled very easily that can be described in the form of a state machine, such as a protocol automaton in a data network.

CTDE domain (Continuous Time / Discrete Event)

The CTDE domain is another time-based domain that is used to describe time-continuous systems, e.g. B. by analog or mixed-signal devices is used. In addition to discrete events, it supports a continuous time course as occurs in real macroscopic systems. The description of systems in the CTDE corresponds to the description of systems with the help of a system of differential equations .

MLDesigner has a number of other domains, such as the HOF domain (High Order Function), which allows a procedural modeling of systems. The BDF domain (Boolean Data Flow) expands the concept of data flow-oriented domains to include the possibility of activating or deactivating model parts based on switching algebraic expressions .

Code generation domains

In addition to all the domains described so far, which are used to model and simulate systems and are therefore referred to as simulation domains, there are a number of code generation domains in MLDesigner, the semantics of which are to be used in the generation of C or VHDL code lies. This makes it possible, in compliance with certain specifications, to change the domain into a code generation domain for a model created in a simulation domain and to generate a desired target code from the same model. The prerequisite for this is that all model elements of the simulation domain must also be present in the code generation domain.

In addition to the use of a code generation domain, it is possible in MLDesigner to generate ANSI-C code, VHDL code or SystemC code directly from models that consist entirely of model elements from the DE domain and the FSM domain.

Areas of application

MLDesigner can be used to design missions, systems, integrated circuits, reconfigurable electronics and a wide variety of other products. It is used to model systems in automobiles, avionics and space systems, but also satellite communication systems and process flows.

literature

  • G. Schorcht, P. Unger, A. George, I. Troxel, D. Zinn, H. Salzwedel, K. Farhangian, CK Mick: System-Level Simulation Modeling with MLDesigner. IEEE / ACM MASCOT 2003 - 11th ACM / IEEE International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems, 12. – 15. October 2003, Orlando, Florida, USA.
  • G. Schorcht: Design of integrated mobile communication systems at mission level. Logos-Verlag, 2000, ISBN 3-89722-462-3 .
  • Herfried Schneider, John A. Buzacott, Thomas Rücker: Operative production planning and control: Concepts and models of information and material flow in complex manufacturing systems. Oldenbourg Wissenschaftsverlag, 2004, ISBN 3-486-57691-7 .

Web links