The ONN.py file provides the sample Python code for the Optimization Assisted by Neural-Networks. The pseudocode, mathematical formulations, benchmark examples and other details are discussed in the paper "Optimization Assisted by Neural Network-Based Machine Learning in Electromagnetic Applications" (currently under review). The code is commented and the variables names correspond to the definitions in the paper.
The Optimization Definitions and Hyperparameters are in lines 38-46. The objective function definition is in lines 61-62.
Once the previous lines of code have been editted accordingly, the user can create a folder under the same directory (the name of the folder is defined in line 30) and simply run the ONN.py code. The code outputs csv files for each optimization in the following format:
iteration variable1 variable2 ... variableN objectivefunction
1 ... ... ... ... ... 2 ... ... ... ... ... ... ... ... ... ... ...
We provided the HFSS slotted patch antenna model. The interested reader can open the file using ANSYS HFSS and take the look at the settings. The model is ready for simulation.