Contents

This document serves as documentation for the COCO (COmparing Continuous Optimizers) software.

This software was mainly developed in the TAO team.

COCO’s website: http://coco.gforge.inria.fr/

Download the pdf version of this documentation.

Contents:

• Introduction
• Installation
  • Requirements
  • Interactive Python
• What is the purpose of COCO?
• First Time Using COCO
  • Running Experiments
  • Post-Processing
  • Write/Compile an Article
• Black-Box Optimization Benchmarking Procedure
  • Symbols, Constants, and Parameters
  • Benchmarking Experiment
  • Time Complexity Experiment
  • Parameter Setting and Tuning of Algorithms
  • Performance Measurement
• Running Experiments with COCO
  • exampleexperiment and exampletiming
  • Testing New Functions
• Post-Processing with COCO
  • Overview of the bbob_pproc Package
  • Standard Use of the bbob_pproc Package
  • Post-Processing from the Python Interpreter
• Python API
  • fgeneric
  • bbobbenchmarks
• bbob_pproc API
  • Post-processing Routines
  • Figures and Tables
  • bbob_pproc.comp2
  • bbob_pproc.compall
  • bbob_pproc
  • bbob_pproc.rungeneric
  • bbob_pproc.rungeneric1
  • bbob_pproc.rungeneric2
  • bbob_pproc.rungenericmany
  • bbob_pproc.genericsettings
• Acknowledgments