Using the BBOB Test Suite on PlatEMO v3.x

To use the BBOB test suite, place the folder in the following path:

PlatEMO\Problems\Multi-objective optimization

This implementation adopts the bbob-biobj test suite. The bbob-biobj suite consists of 55 test problems, each characterized by 15 different parameter configurations and 6 alternative numbers of variables (i.e., 2, 3, 5, 10, 20, and 40).

---

How to Build the BBOB Test Suite for MATLAB from Scratch

Follow these steps:

1. Download COCO v2.6.3 (GitHub repository)

2. Configure COCO Follow the instructions in the README file provided on the homepage of COCO v2.6.3 GitHub repository.

   • Compiler Setup: For example, use MinGW for compilation. Ensure that the system environment variables are properly configured; otherwise, MATLAB will not recognize the compiler.
   • MATLAB Compiler Check: Verify that MATLAB is configured with a C language compiler by running the following command in MATLAB:

     mex -setup

     If issues arise, refer to relevant tutorials for guidance.

3. No Need to Install cocopp

4. Set Up a Python Environment

   • Create a virtual environment using Conda (Python v3.6).
   • Run the following command:

     python do.py run-matlab

     This command will automatically start MATLAB and close it twice.
     Note: Any Python interpreter compatible with the version should work, as no additional Python packages or dependencies are required to execute do.py.

5. Extract the Required Folder
   After completing the configuration, locate the following folder:

   ..\coco-2.6.3\code-experiments\build\matlab
   

   This folder contains all the necessary files. You can rename it if needed (e.g., coco_matlab).

---

Additional References:

• tpb (GitHub repository)