摘要 :
We present Test Interaction Console (tico), a toolbox for MATLAB to assist the system developer with authoring and executing large parameterizable test suites. It can be considered as an add-on layer to the core testing framework ...
展开
We present Test Interaction Console (tico), a toolbox for MATLAB to assist the system developer with authoring and executing large parameterizable test suites. It can be considered as an add-on layer to the core testing framework in MATLAB, where our focus is on ease of use and providing high-level tools to automate recurring tasks in the aerospace domain of model-based design testing, tico is built around a custom command line dispatcher that mimics Bash syntax and has a flexible plugin architecture to extend its capabilities. In the toolbox, we provide enough plugins to connect MATLAB test artifacts with the DevOps platform GitLab and Polarion servers as an example for application life-cycle management (ALM) environments. This integration is the foundation for a complete round-trip solution to validate high-level aircraft requirements defined in Polarion and their test cases executed in MATLAB. Our custom 'grid' data container simplifies the parameterization of test cases over flight envelopes with thousands of iterations and encourages a style of dataflow programming. Tag annotations of file artifacts on test cases enable tico to resolve a graph of task dependencies and speed up subsequent tasks by reusing data files across the test suite. All configuration for the toolbox is handled with environment variables, simplifying execution and deployment on virtual platforms, such as Docker.
收起
摘要 :
We present an open, modular toolbox for Matlab that implements an algorithm for Subset Simulation (SuS) with Markov-chain Monte Carlo (MCMC) sampling to estimate occurrence probabilities of rare events, available at http://go.tum....
展开
We present an open, modular toolbox for Matlab that implements an algorithm for Subset Simulation (SuS) with Markov-chain Monte Carlo (MCMC) sampling to estimate occurrence probabilities of rare events, available at http://go.tum.de/605579. The toolbox is built around an object-oriented API that makes understanding the setup easy for the user. Core features include four different MCMC sampling methods, three different limit-state evaluation modes, accelerating slow limit-state function evaluations through surrogate models and through compilation of Simulink models, reusing result data to study parameter sensitivities, sampling dependent parameters from Vine Copula structures, and an instrumentation approach to tracking algorithm progress. We demonstrate in short examples how we use these features to easily analyse the rare-event failure probability and parameter sensitivities of benchmark problems, and for a performance requirement on a closed-loop aircraft simulation model.
收起
