3. TEXT
Introduction to numerical optimization using DAKOTA and OpenFOAM®
Instructor: Joel Guerrero
Summary: In this training session the attendees will be introduced to numerical optimization using
DAKOTA. We will address how to conduct parametrical and design of experiments studies, single and
multi-objective optimization, surrogate based optimization, exploratory data analysis, and how to
couple DAKOTA with OpenFOAM® (or any other black box solver). To follow this training session a
basic knowledge of OpenFOAM® is required. No prior knowledge of DAKOTA is required.
Abstract: Course_abstract-Joel_Guerrero-Dakota.pdf
Training type: Intermediate
Session type: Undefined
Software stack: OpenFOAM 3.0.x, Dakota 6.3, OpenSCAD, Salome 7.7.1, Python 2.7 (Anaconda).
Training Material: https://drive.google.com/open?id=0BwfuSMqexhZxUExmbTF4Z2Ztckk
8. Optimization method selection
Method Variable Constraint
Gradient Based
Local
Smooth
Continuous
No bound
Bound
Linear & nonlinear
Gradient Based
Global
Smooth
Continuous
Bound
Linear & nonlinear
Derivative Free
Local
Nonsmooth
Continuous/
Decreate
Bound
Linear & nonlinear
Derivative Free
Global
Nonsmooth
Continuous/
Decreate
Bound
Linear & nonlinear
12. Dakota_of_arhmed_multi.in
(つづき)
interface
fork
asynchronous
evaluation_concurrency = 2
analysis_driver = 'simulator_script'
parameters_file = 'params.in'
results_file = 'results.out’
work_directory directory_tag
copy_files = ‘templatedir/*’
# uncomment to leave params.in and results.out files in work_dir subdirectories
named ‘workdir‘ file_save directory_save
aprepro
## when using conmin_frcg (above) with analytic_gradients (below),
## need to turn off the active set vector as the interface does not parse it.
deactivate active_set_vector
計算実行命令ファイル
パラメータ入力ファイル
計算結果出力ファイル
テンプレートファイル
外部ソルバー利用
複数条件の同時計算
同時計算数
ワーキングディレクトリにタグをつける 例 条件番号Nのときworkdir.N
ワーキングディレクトリ名 ファイル保存ON ディレクトリ保存ON