0.8-beta

New Features

• the displays provided by the post-processing have improved significantly, including adjusted target values

Bug Fixes

• quite a few

0.7-beta.6

Bug Fixes

• post-processing install failed on Windows

0.7-beta.4

Bug Fixes

• Python cocoex module (experiments in Python): the bi-objective functions returned a reference to a numpy array which changed in subsequent calls.

Changes

• Added a few missing functions to the Matlab interface.
• Figures display different targets, not the final targets yet.
• ppdata.html file as entry point to browse the results

0.7-beta.3

Bug fixes

• erroneous endif in Matlab wrapper replaced by end (thanks to Abdullah Al-Dujaili for pointing this out)

Changes and new features

• minor improvements and extensions in the Matlab/Octave wrapper
• documentation sources updated (without yet pushing it to the web)

Known issues

• postprocessing not yet finalized; results should therefore be interpreted with precaution

0.7-beta.2

Bug fixes

• first target is not fixed anymore
• several smaller bugfixes, in particular in postprocessing

Changes and new features

• new lean Matlab/Octave interface, allowing for output in Matlab and access to cocoSetLogLevel; compilation now >10 times faster
• Octave under Windows is not interfering with run-c anymore due to new octave_coco.bat file
• online documentation of Coco functionality within Matlab/Octave (via help and doc)
• implementation of restarts in example experiment for C, python, and Matlab/Octave
• several improvements of postprocessing

Known issues

• postprocessing not yet finalized; results should therefore be interpreted with precaution

0.7-beta.1

small bug fix in testing code

0.7-beta

Changes

• Octave interface is more stable
• C++ works using the C example code
• small change in the Python experimental interface of the Suite
• nicer output when the experiment is running
• different verbosity levels available (log_level)
• many smaller and larger bug-fixes

Known Issues

• example experiments need some more polishing, some don't implement restarts yet
• figures and tables in the bi-objective are still provisional

0.5-beta.3

• bug-fix when asserting the optimal f-value, which could lead to false positive outcomes and consequently termination.
• pagination in the Matlab/Octave example experiment is now turned off (more off;), as it should be.
• known issues
  • in some cases, Octave doesn't build under Windows.
  • figures and tables from the post-processing for bi-objective data are expected to contain bugs and glitches

0.5-beta.2

Changes

• Bugfix for the Matlab example experiment.
• Python example experiment runs now only one batch by default.
• Minor adaptation of pathnames.

0.5-beta.1

Fixed a problem which prevented the post-processing to run in bi-objective data.