Hacker News Reader

Ansys, Abaqus CAE, Dyna, Optistruct, FEKO, AcuSolve, NASTRAN, ... The list goes on and on. Each bigger DE or partial DE has a huge ecosystem of numerical software around it to approximate solutions for real world applications.

High and low electro magnetics (Maxwell), Navir-Stokes, ...

View on HN

What tools are used for parameter estimation? E.g. I made a model that describes my system behavior, then collected a lot of data and next need to find the constants in the model that make the best fit. Thank you.

The search terms there would be "Model Calibration" and/or "System Identification" Software. I think those are again very domain specific. But there are a few. PS: While AI and machine learning are catching on in many different fields, the software packages in those domains make heavy use of fairly complicated fitting and approximations methods already. Lots or correlation and effect estimation, which builds many of the underlying guts of machine learning.

Engineers should be taught about actual numerical algorithms and their qualitative properties. Cases in which there exist explicit symplectic algorithms, etc.

Most engineering programs do have at least one full semester of numerical methods, even if in practice most engineers use matlab, scipy, abaqus etc...

Reply on news.ycombinator.com