A large-scale system modeling environment has been released by Wolfram. SystemModeler is intended as a very general environment that handles modeling of a wide range of systems.
A large-scale system modeling environment has been released by Wolfram. SystemModeler is intended as a very general environment that handles modeling of systems with mechanical, electrical, thermal, chemical, biological, and other components, as well as combinations of different types of components.
On his blog Stephen Wolfram says that the problem with existing modeling environments is that they are limited to producing block diagrams that poorly represent key components, so the models that are created are limited to being used only for simulation, not engineering analysis or integrated computations.
The idea behind SystemModeler is that you can put together a high-fidelity model that can be used for everything from optimization and control design to custom visualization and interactivity. It can also be used together with Mathematica to give you a method for complete programmatic control of simulations. You can also use Wolfram Alpha to bring in real-world data for use with the model.
As in Wolfram’s Mathematica, everything in SystemModeler is represented in symbolic form. To create a model you put together a hierarchy of connected components that can be assembled interactively by dragging and dropping. SystemModeler comes with a library of standard components such as sensors, actuators, gears, resistors, joints, heaters, and so on. You can also make use of model libraries from component manufacturers because SystemModeler uses the standard Modelica language for system specifications.
Behind the scenes SystemModeler then interprets the symbolic system description and derives a collection of differential-algebraic and other equations and event specifications, which it then solves to produce a fully computable representation of the system. The representation mirrors what an actual physical version of the system would do, but allows instant visualization, simulation and analysis.
The blog post about SystemModeler has an example of SystemModeler in action showing a 2,685-equation dynamic model of an airplane being used to analyze the control loop for continuous descent landings.
Microsoft's Project Malmo platform for AI experimentation is now available on Github. The Malmo platform is built on top of Minecraft and uses its experiences and interactions to learn how to interact [ ... ]