CamelCMA/DOTWrapper

Python wrapper for the DOT optimizer

About

This repository is forked from MODRG/DOTWrapper, created by Prof Gerhard Venter on 21 August 2019. The user is recommended to go through its README file first.

Requirements

• The DOT optimizer itself (commercially available form VR&D). A Python wrapper for DOT is provided that will allow the user to call the DOT shared library from Python. It should be noted DOT is a commercial optimizer, a license is required to use it.
• NumPy & Matplotlib Python library

File description

File	Description
dot.py	The Python wrapper for DOT
dot.pdf	The DOT manual
sec4_2_BoxDesign.py	The first example problem from the DOT manual, using the provided wrapper

Figure for model 1

Figure for model 2

Figure for model 3

Modification

Reconstruct the dot.py file and implement the following methods to call:

• print_init can be used to check the input parameters.
• fit calls DOT to resolve the problem.
• print_info prints the following values to screen.
  • Objective function value
  • Max violated constraint value
  • Design variable values
• plot_fig plots the following values as figures. A flag is_savefig=True can be used to save the figures and a flag dpi can be used to set the resolution of the figures.
  • Objective function value
  • Max violated constraint value
  • Design variable values

Problem-defined steps

4 steps to complete the setting for a optimization problem:

1. Define the problem in myEvaluate function.
2. Define nDvar & nCons as integer.
3. Define x, xl, xu as numpy ndarray.
4. Configure the model(s) with corresponding method(s) to call.

Typically, the user is recommended to explore the solution by setting different nMethod value. Sometimes, interesting outcomes could be observed.

Acknowledgement

• Thanks for Prof Gerhard Venter constructing the original dot.py file.
• Thanks for VR&D sponsoring the trial license of DOT.