byADIT Free software for engineers and students
EngiFree
Dakota — The Complete Guide: Optimisation and Uncertainty Quantification Around Any Simulation Code

Dakota — The Complete Guide: Optimisation and Uncertainty Quantification Around Any Simulation Code

Dakota is an open-source toolkit (LGPL-2.1) from Sandia National Laboratories. The idea is simple: take an existing computational model — finite elements, CFD or any other program — and Dakota runs it repeatedly with varying inputs and analyses the results.

What you get

How it works

Dakota connects to the simulation program through input and output files, so the program itself does not need to change. The learning curve is steep and assumes familiarity with statistical methods. The prebuilt Linux binaries are built for RHEL, and on other distributions you may need to compile from source.

Related tools

For multidisciplinary optimisation in Python — OpenMDAO. For uncertainty quantification and reliability analysis in Python — OpenTURNS. For calibration and parameter estimation, common in groundwater models — PEST++.

The bottom line

Dakota turns any simulation into a design and reliability tool — free for any use, including commercial work, for those willing to invest in learning it.

Further reading

מדריכי AI למהנדסים ולמשרד, ב-5 שפות: adit-ai.com ←