The repository contains a HyperStudy project for driving the optimization process and two python files for computing and linking with the HyperStudy.
The HyperStudy project already contains results for two optimization processes, one with a genetic algorithm and the other with a multiobjective search. They both have around 4000 iterations.
Open the two_optimizations.hstudy in Altair HyperStudy. The setup will be out of date but the two runs should be valid. Optimization_genetic_4300i contains the results for the genetic algorithm and the Optimization_search contains the results for the multiobjective search.
In order to run it there are some prerequisite steps that must be completed:
- Install (if not already installed) Python 3.8 or later;
- Open the project in HyperStudy;
- Go to Edit -> Solver Scripts and change the path to python.exe for the script "Python_VEnv"; this lets HyperStudy know what executable to run on each iteration;
- Go to Setup -> Define Models -> Solver Input Arguments and change the arguments to "path_to_hyperstudy.py ${file}" (example: "C:\repo\hyperstudy.py ${file}"; this step ensures that the python file is called with the input file as the first argument.
Now the setup is prepared for testing the model and then optimizing it.
The python script expects an input file with arguments for the model. On each line, the file contains: the shell thickness, the core thickness, the plate length, the shell material and the core material. File example:
0.01000
0.09000
3.50000
Aluminium
DivinycellH100
The python scripts solves the model with these parameters and writes in output: dispalcement, maximum displacement, core stress, maximum core stress, shell stress, maximum shell stress, cost of materials and total mass. Example of output file:
12013148688.75901
294.19950
29.58165
0.24000
10296982.50000
40.00000
0.22
0.11025
In HyperStudy in the "Define Input Variables" and "Define Output Responses" these parameters are defined and optimization goals are added.