Process automation and optimization
The interaction of different software packages and the exchange of data are essential in today's complex workflows. Since the susceptibility to errors increases with each manual input and automation via script and subroutine development and the maintenance of interfaces is time-consuming, Dassault Systèmes provides Isight, a product that independently organizes communication in design processes. From simple parameterized optimization loops to fully automated result evaluation - Isight adds significant value to your simulations.
With "drag and drop", even complex flowcharts can be created quickly and easily. A wide range of process and application components are available for creating so-called Sim-flows. The application components include, for example, the interfaces to Abaqus, CATIA and Dymola. In addition to Dassault Systèmes products, interfaces to Excel, Word and MATLAB are also provided, to name just a few of the common applications outside the Dassault portfolio. In this way, Isight combines the functionality of a wide range of applications in your company into cross-program design processes.
The DOE component provides a variety of algorithms to representatively select process parameters from a defined design space and automatically provide them in the respective software. The influence of the parameters is not separated, but is considered under mutual influence. Predefined target variables and target functions determine the best solution from the selected parameter values.
This component enables a parameterized system optimization based on a wide range of optimization algorithms such as gradient methods, multi-island genetic and multi-objective particle swarm algorithms. The right choice of the optimization algorithm can have a great influence on the calculation time and quality, therefore Isight also provides a Pointer Automatic Optimizer that automatically determines the best algorithm.
Monte Carlo and Six Sigma
The Monte Carlo and Six Sigma components provide statistical methods for determining system sensitivity. This can be done by generating random variables around a mean value or by using a statistical sampling procedure. Evaluations of the parameter sensitivities lead to a better understanding of the process and to possible improvements.