Exercises

In order to optimally follow the practical exercise classes offered during the summer school, we request the participants to bring their own laptop.

To not reduce the bandwidth of the wireless internet access in Quiberon substantially, we furthermore ask all participants to install the following software packages and download the following data files before the summer school.

Software to install for the exercises

Numercial Blackbox Optimization

• Octave or Matlab
• Exercise sheet
• cmaes.m
• plotcmaesdat.m

Combinatorial Optimization

• Presentation with correction (<1MB)
• Exercise -- files to download (<1MB)
• ParadisEO
  • ParadisEO requires C++, cmake, doxygen and gnuplot to be installed
  • ParadisEO can be installed under Linux, MacOS and Windows

Performance Evaluation of Anytime Blackbox Optimizers

• install ipython+numpy+matplotlib (compare https://www.lri.fr/~hansen/cmaes_inmatlab.html#python-remarks), preferably and most easily via
  
  http://continuum.io/downloads.html
  
  (then "ipython notebook --pylab inline" should open an "ipython notebook" in the default browser)
  
  
• download https://transfert.inria.fr/fichiers/64ef3c7675e54ebeffd0e92de9b0307d/performance-evaluation.tar.gz (98MB)

• helpers.py (< 1MB)

• performancperformance-evaluation-session.ipynb(IPython notebook file, < 1MB)
  When IPython is started in the folder containing this file, the browser content from the session is reloaded (the respective python code in the cells needs to be executed again). A purged command history is given below, which should work from a "fresh" IPython shell, given the folders bbob.v13.09 and rawdata are available.

Parallel Evolutionary Algorithms

• Download easea (3 MB)
• onemax_template.ez (< 1 MB)