BayesOpt v0.6 released.

[Ruben Martinez-Cantin]

With this update, there is not much changes on the outside, but the
inside have changed dramatically. Now, the library is under heavy
development to add new algorithms and include more robust techniques
for the ones already implemented.

If you have trouble compiling after the update, try cleaning the CMake
cache (or delete CMakeCache.txt). Also the interface have slightly
changed the learning parameter is incompatible with the previous one.
If you were not modifying that parameter in your code, then it should
be fine. More info in the documentation:
http://rmcantin.bitbucket.org/html/

Changelog:

-Complete refactoring of inner parts of the library. The code is
easier to understand/modify and it allow simpler integration with new
algorithms.

-Updated to the latest version of NLOPT (2.4.1). Wrapper code symplified.

-Error codes replaced with exceptions in C++ interface. Library is
exception safe.

-API modified to support new learning methods for kernel
hyperparameters (e.g: MCMC). Warning: config parameters about learning
have changed. Code using previous versions might not work. Some of the
learning methods (like MCMC) are not yet implemented.

-Added configuration of random numbers (can be fixed for debugging).
Fixed issue with random numbers using different sources or random
number with potential correlations. Now all the elements are
guaranteed to use the same instance of the random engine.

-Improved numerical results (e.g.: hyperparameter optimization is done
in log space)

-More examples and tests.

-Fixed bugs.

-The number of inner iterations have been increased by default, so
overall optimization time using default configuration might be slower,
but with improved results.

Happy coding!

Ruben