Column-Design-Optimization

Requirements:

1. OS Windows/Linux/Mac.
2. Anaconda
3. Python>3.8

How to use?

1: Setting working environment

Create an environment using the env.yaml in Anaconda with the following command:

conda env create -f env.yaml

2: Data generation

The data generation folder contains files for parametric data generation.

• main.py is the main script to run the data generation.

• column.py script contrains the material, geometric, analysis, and model parameters.

• functions.py contains function used to generate data, such as section analysis, random generation of section geometry.

• data.csv file contains the sample output from data generation.

3: Data pre-processing and network model

• pre-processing.py script contains the data filtration based on monetary cost for the case of 4.0 meters.

• normalization.py script contrains min-max normalization and preparing the data for network.

• net.py file contrains network file for training the network.

• use column.pth and test.py files for inference.

• use test_min.h5 and test_max.h5 files for minmax normalization of the test sample before extracting network predictions.

4: Design check

Use check.py to run the check of the network output results.

5: Publication: https://www.isec-society.org/ISEC_PRESS/EURO_MED_SEC_04/xml/SUS-14.xml